3:20 PM -- I listened to a very interesting talk on defensive and offensive binary development the other day, which was led by one of the lesser known but very tech-savvy guys at Symantec's research group in Austin. The session addressed well-known binary armoring techniques, but it also raised an interesting issue that hasn't been discussed.
Often, we take it for granted that we can keep dangerous code in a sandbox. We can use proxies to hide our IP addresses, we can spoof email to hide our real identities, and we can sandbox an environment with VMware to protect our machines while we look at potentially malicious binaries, right?
Unfortunately, that's no longer the case. It may be old news to the vast majority of malware researchers out there, but over the last few years, malware writers have given their code the ability to detect if it is inside a VMware session. This allows the worm writers to find out if someone is attempting to analyze their exploits.
This capability puts a big crimp in a lot of anti-malware detection services that attempt to do large-scale sweeps of the Internet in hopes of finding malware signatures before their competitors do.
This kind of defensive malware authoring isn't unique -- there are lots of techniques to prevent detection. But I believe we'll be seeing a lot more armoring techniques in the future, which will make detection of zero-day malware more and more difficult.