Thursday, March 4, 2010

RAM Analysis Part 2

Back in January of this year, Harlan posted on his blog some information about a concept called Least Frequency of Occurrence (LFO). Basically, this concept established by Pete Silberman states that when performing analysis containing potentially malicious binaries or dlls, to look for small variations of "normal" versus the huge spikes in activity one might think is associated with malware activity.

So in a case I was working on, I was able to put this theory to the test. I had a RAM dump from what was believed to be an infected system. I then used Memoryze to process the dump and Audit Viewer to review the data.

So what would you look for? Well, like I have said many times, you cannot approach any investigation with some sort of idea of what you are looking for. You cannot expect to fire up a RAM dump and simply spot the anomaly. It's not realistic and will likely never ever happen. So in this case, I had an idea of what I was looking for. Before opening the RAM dump, I was able to review the volatile data I had extracted from the suspect system, and found a series of files with MACB times around the time the incident was reported to have occurred. Now, that in an of itself is a huge issue which is outside the scope of this to our story.

With my list of files that I was looking for (which was about 15 files), I opened the RAM dump by reviewing the dlls and sorted each by the frequency. And after a bit of searching, I found the one file I was looking for injected into IEXPLORE.exe.

So, this is a great example of how what I was looking for and found was not a huge spike in data point. So now, I have the name of the injected dll, and the PID it was injected into. Now, that might be such a big deal with a process like IEXPLORE.exe, but, what if the process was something a bit more populous on the system like svchost, or lsass? Then having the specific PID is very important.

So next, I could go to my forensic image, extract the dll and do some reversing to figure out what this thing is doing and why it's there.

Good times!

No comments:

Post a Comment