Firms Far From Taming The Tower Of APT Babel
Threat intelligence firms continue to have individual lexicons for advanced persistent threats, making information sharing more difficult
Companies looking to learn more about the attackers who are targeting their systems will likely have to translate between reports from different incident-response and security firms.
While early efforts are under way to allow threat-intelligence firms to better exchange information on adversaries and their tools, security companies each have its own naming scheme, making identifying common threats more difficult. Take the Comment Crew, for example. The espionage group, which targets intellectual property in nearly a dozen industry sectors and is thought to be connected to an intelligence component of the People's Liberation Army, is also known as Comment Panda by security startup CrowdStrike and APT-1 by incident-response firm Mandiant. Managed security service provider Dell Secureworks puts the group into a broad bucket known as the Shanghai Group, based on its infrastructure's location.
More Security Insights
- Forrester Study: The Total Economic Impact of VMware View
- Securing Executives and Highly Sensitive Documents of Corporations Globally
- Top Big Data Security Tips and Ultimate Protection for Enterprise Data
- Client Windows Migration: Expert Tips for Application Readiness
"Different companies are approaching the same threats from different angles," says Joe Stewart, director of malware research at Dell Secureworks. "I focus on the malware and the IP infrastructure used by the attackers, whereas a response company sees different tools, and that leads them to a different naming scheme."
Like the proverbial blind men describing the elephant, each security firm collects intelligence on specific aspects of targeted attacks, and so tends to see different parts of the overall threat. The result is that different companies tend to group attacks into categories that do not appear to match.
CrowdStrike, for example, uses a cryptonym system, coming up with code names for the actors based on an umbrella term for the type of actor: Panda for China, Kitten for Iran, and Spider for certain criminal groups, for example. The company then adds a second identifier onto the more general name. Anchor Panda is the designation for a group thought to be connected to the Chinese PLA navy.
Mandiant, on the other hand, started with a single designation, APT1, and expanded its lexicon, adding APT2 and APT3 as new espionage groups were discovered. The company has approximately two to three dozen groups it currently tracks.
The problems are apparent when comparing names for another espionage group, which favors a tactic of dynamically calculating the port used for command-and-control communications. It is known as Numbered Panda by CrowdStrike, APT-12 by Mandiant, and by a variety of other common names, such as DynCalc, Ixeshe, and JoyRAT by other security firms.
[Researcher uncovers hundreds of different custom malware families used by cyberspies -- and discovers an Asian security company conducting cyberespionage. See Scope Of APTs More Widespread Than Thought.]
The problem makes it more difficult for threat-intelligence firms to share information with each other about threats, says Adam Meyers, vice president of intelligence for CrowdStrike.
"We've seen a problem in that there is no common lexicon for how we describe targeted attacks," Meyers says. "We are quickly turning into the antivirus industry."
Antivirus companies are well-known for independently choosing names to describe the same malicious program, a problem that has often caused confusion among consumers. In 2005, the MITRE Corp. created an initiative to find common names for malware, but closed down the project in late 2006. Instead, the government contractor began developing a common language to describe malware characteristics, known as the Malware Attribute Enumeration and Characterization (MAEC).
Researchers at the threat-intelligence firms have started informal discussions on similar efforts for describing and comparing attributes of current threats. The companies hope that finding a common lexicon, or finding ways to translate their lexicons, can result in the better identification of threats.
"We think that getting the industry as a whole to share some of this intelligence and to communicate with the same lexicon is going to allow us to be more effective in helping our customers," CrowdStrike's Meyers says.
Yet the companies still need to hash out many of the basic details, such as differentiating the attack campaigns from the attack groups, and determining how much of their proprietary information will be revealed. Mandiant, for example, has focused on releasing indicators of compromise -- specific clues that can help companies identify whether a certain group is in their network. In its APT-1 report released in February, the company released a large number of indicators of compromise that can identify if the group has attacked a network.
A significant problem in sharing the information on threats is that the evidence is tied closely to the way a security firm conducts its investigation, says Aaron Cherrington, a member of the intel team at Mandiant.
"It is one of those things where what you actually observe from your perspective as a security company really defines how you create your groups," he says. "And what information are you collecting allows you to make more of those links between attacks and the attacker."
In most cases, that information is proprietary and sensitive, which makes sharing the information that much more difficult, he says.
Have a comment on this story? Please click "Add Your Comment" below. If you'd like to contact Dark Reading's editors directly, send us a message.