The first Mandiant Incident Response Conference (MIRCon) is now in the bag, so to speak. It was an impressively valuable and fun-filled two days, and I have to thank Mandiant once again for throwing down on an excellent shindig. As with my review of Day 1, I'll recap some highlights from the various presentations. Those of you who weren't able to attend may also be interested in the recap webinar that Mandiant is presenting next week (Oct. 19): State of the Hack: The Hangover - What REALLY happened at MIRCon.
The Day 2 keynote was delivered by Gordon Snow, Assistant Director of the FBI's Cyber Division, who spoke about the FBI's perspective on "cyber threats." Mr. Snow discussed the range of threats and attackers the FBI deals with, and explained how the FBI's approach to computer-based crime has evolved over the years. This evolution has moved the FBI to a far more proactive role in Cyber Defense, as they now make an effort to notify organizations who may be targeted by attackers they are currently investigating. This approach arose, in part, from the understanding that:
"Cyber information is unlike any other kind of information. It's perishable. If I don't get it to you in a reasonable period of time, it's useless to you." - Gordon Snow, Assistant Director, FBI Cyber Division
What Mr. Snow called "cyber information" can more accurately be called threat intelligence in this context. But he's spot-on concerning the perishable nature of threat information. He went on to point out that the FBI's ability to contact an organization and share information is partly dependent on how easy it is to identify an appropriate contact and whether a trust relationship already exists. He encourages security and incident response professionals to develop relationships with their local FBI field offices before a problem arises.
This was followed by a brilliant presentation from Thomas Dullien (Halvar Flake) entitled Creating Information Asymmetry to Ambush Attackers. He posited that the fundamental problem with signature based defense technologies (and antivirus in particular) is that they reveal information to the attackers about what we (as defenders) know and what we've detected. The signatures become a source of information for the attacker, who can then refine his or her attacks to bypass current defenses. As a result, the attackers always have more information about us (the defenders) than we have about them. He then discussed some analysis techniques and tools that zynamics has created to overcome this problem. And, from the sound of things, we'll see more of this also making its way into Mandiant's products, as they are already utilizing VxWorks. I can't do justice here to the ideas and technologies Thomas presented, so I strongly encourage you all to read some of his recent blog posts that address the same issues:
- Challenging Conventional Wisdom on AV Signatures (Part 1 of 2)
- Challenging Conventional Wisdom on AV Signatures (Part 2 of 2)
Following on this theme, Kevin Mandia and Richard Bejtlich held and impromptu panel discussion before lunch, in order to address some questions that time did not permit on Day 1. One of the the recurring questions that arose was whether or not to submit malware samples to antivirus vendors. For exactly the reason Thomas Dullien presented, there is danger of both giving away intelligence and compromising your own remediation efforts during Incident Response. The general consensus was that advanced, targeted malware should not be shared with AV vendors, because the value represented by that intelligence (the ability to detect the intruder without tipping them off) is far outweighed by having a new antivirus signature that will simply cause the intruder to adjust. On the other hand, there may be less risk/harm in sharing mass-malware samples with AV vendors.
For a lighter look at malware, Peter Silberman and Nick Harbour presented The Malies ™ during lunch. In case you don't know, The Malies is "an Award Show for Epic Fail and Great Success in Malicious Software." Since I believe they will be presenting this again elsewhere in the near future, I won't steal their thunder by revealing results here. But just to give you a small taste of the fun, a few of the award categories were "Most Perplexing Misuses of the English Language," "Most Blatant Disregard for Getting Caught," and "Ligatt Security Award for Plagiarism."
This was followed by The Nostradamus Panel, where Kevin Mandia, Amit Yoran and Byron Collie discussed the future of Incident Response. I was a bit disappointed here, unfortunately. The panel seemed to derail a bit into questions such as whether breach notifications provide any value, and didn't spend enough time on the future of Incident Response. But if you want more detail about this, Mandiant is offering a webinar next week in which they'll provide their own recap of MIRCon.
The day closed on a surprisingly excellent presentation called Legal Jujitsu: The Art of Avoiding the Pain Caused by a Breach, presented by Lisa Branco and Jake Sommer from Zwillinger Genetski, LLP. While no one audibly groaned (that I noticed), you can feel the collective "Ugh!" in the room when a legal presentation is introduced to a room full of geeks. But the presenters were energetic and engaging. And they covered some good material regarding what lawyers focus on during Incident Response, and how they can save you (as a responder) and your organization a great deal of pain when properly engaged in the response process.
They shared far more than I can pass on in this space, but some of the more interesting nuggets were those that explained a lawyer's role in the Incident Response process. First, they identified the lawyer's #1 goal, which is to minimize the liability resulting from a data breach, then gave some specifics about the lawyer's role:
- Analyze disclosure obligations and data security regulations to ensure compliance.
- Preserve privilege where possible. (Use of client-attorney privilege can help to protect sensitive data.)
- Help ensure investigation continues in the right direction. (This includes keeping everyone focused on answering those questions that have legal significance, such as "What data was compromised?" and "What was damaged?")
- Interface with 3rd-parties so that key internal personnel can continue to do their work. (In other words, having a good lawyer on your side can be like saying "Leave me alone so I can forensicate!")
- Manage communications with various audiences (such as Law Enforcement, business partners, regulators, the public, etc.)
- Ensure effective defense by preventing company from making damaging external or internal statements.
From a responder perspective, I was particularly intrigued by their stance that the lawyer should serve as the communication channel between the technical incident responders and management. This, of course, requires a lawyer who is technically savvy enough to serve as translator and convey the correct information to both sides. Where such a lawyer is available, I think this is a great idea. But gaining that particular expertise, which is often beyond the capacity of General Counsel, would require many organizations to hire or contract lawyers with specific expertise in handling data breach issues. That's probably still money well spent in many cases, but may not always be feasible.
So that wraps it up for me and for MIRCon. If you presented at, attended, or simply wanted to attend MIRCon, I'd love to get your feedback in the comment section. Thanks again to Mandiant for putting together an excellent event and making it available to the community at such a reasonable cost.
Gregory Pendergast is Interim Information Security Officer at Virginia Commonwealth University.