The Software Similarity Problem in Malware Analysis

In software engineering contexts software may be compared for similarity in order to detect duplicate code that indicates poor design, and to reconstruct evolution history. Malicious software, being nothing other than a particular type of software, can also be compared for similarity in order to detect commonalities and evolution history. This paper provides a brief introduction to the issue of measuring similarity between malicious programs, and how evolution is known to occur in the area. It then uses this review to try to draw lines that connect research in software engineering (e.g., on "clone detection") to problems in anti-malware research.