Assembly Code Clone Detection for Malware Binaries

Assembly Code Clone Detection for Malware Binaries Mohammad Reza Farhadi Malware, such as a virus or trojan horse, refers to software designed specifically to gain unauthorized access to a computer system and perform malicious activities. To analyze a piece of malware, one may employ a reverse engineering approach to perform an in-depth analysis on the assembly code of a malware. Yet, the reverse engineering process is tedious and time consuming. One way to speed up the analysis process is to compare the disassembled malware with some previously analyzed malware, identify the similar functions in the assembly code, and transfer the comments from the previously analyzed software to the new malware. The challenge is how to efficiently identify the similar code fragments (i.e., clones) from a large repository of assembly code. In this thesis, an assembly code clone detection system is presented. Its performance is evaluated in terms of accuracy, efficiency, scalability, and feasibility of finding clones on assembly code decompiled from real-life malware binary files and some DLL files from an Operating System. Experimental results suggest that the proposed clone detection algorithm is effective. This system can be used as the basis of future development of assembly code clone detection.