Deriving common malware behavior through graph clustering
Younghee Park, Douglas S. Reeves · 2011
Detection of malicious software (malware) continues to be a problem as hackers devise new ways to evade available methods. The proliferation of malware and malware variants requires methods that are both powerful, and fast to execute. This paper proposes a method to derive the common execution behavior of a family of malware instances. For each instance, a graph is constructed that represents kernel objects and their attributes, based on system call traces. The method combines these graphs to develop a supergraph for the family. This supergraph contains a subgraph, called the HotPath, which is observed during the execution of all the malware instances. The proposed method is scalable, identifies previously-unseen malware instances, shows high malware detection rates, and false positive rates close to 0%.