Fragile Watermarking For Hardware IP Blocks

Abstract

In this chapter, we surf over different techniques of Hardware Trojan detection and focus on the strength and weakness points of each in order to manifest the cons and pros of our proposed fragile watermarking design. We display some important metrics that must be achieved by a reliable Trojan detection technique; we then test our proposed watermarking approach against each in order to support our idea about fragile watermarking serving as an effective solution to a major security attack like that of Hardware Trojans. We then conduct a comparison between our proposed fragile watermarking approach and other detection methods based on a set of performance evaluation criteria including insertion time, repeatability, implement-ability, coverage scope, success chance, and infrastructure needed. Our proposed tool is proved to have the least impact on the design cycle’s overhead and a low complexity level as well. Simulation results are presented to show the insignificant effect of our watermarking design on the delay time of the original system; this was represented by the small insertion times in comparison to the increasing number of generated watermark bits. Finally, a testing framework is introduced to simulate a number of Trojan attacks and evaluate the sensitivity of our watermarked design towards such manipulations. Our proposed watermark is shown to break once the least alteration occurs and hence is efficient in securing various designs from intentional manipulations.