Towards Enhancing Power-Analysis Attack Resilience for Logic Locking Techniques

Logic locking techniques have been widely investigated to thwart intellectual property (IP) piracy and reverse engineering attacks on integrated circuits. Although extensive research efforts have been made to examine the resilience of logic locking techniques against Boolean satisfiability (SAT) and key sensitization attacks, there still lacks a comprehensive assessment of different locking methods' resilience against power-analysis attacks. In this work, we evaluate the success rate of differential power analysis (DPA) and correlation power analysis (CPA) attacks that are performed on the circuits encrypted with logic locking techniques applied at gate level or transistor level. To enhance the CPA attack resilience of the existing transistor-level locking techniques, we further propose a new strategy to search for optimal key insertion locations. Our analysis and experimental results indicate that gate-level locking and transistor-level locking should use different strategies to select the optimal key insertion locations. Our case studies confirm that the proposed key insertion strategy can improve the transistor-level locking technique's resilience against CPA attacks.