Software Security through Targeted Diversification

Despite current software protection techniques, applications are still analysed, tampered with, and abused on a large scale. Crackers compensate for each new protection technique by adapting their analysis and tampering tools. This paper presents a low-cost mechanism to effectively protect software against global tampering attacks. By introducing diversity per programme instance, we illustrate how to defeat various patching methods using inlined code snippets. We propose an efficient technique for creating the snippets based on genetic programming ideas, and illustrate how our approach might trigger a small-scale arms race between defending and attacking parties, each forced to evolve in order to “stay in the game”.