Algorithmic collision analysis for evaluating cryptographic systems and side-channel attacks

Side-channel attacks have emerged as a kind of effective security threat targeting system implementation of cryptographic algorithms. Evaluating a cryptographic system's resilience to side-channel attacks is therefore important for secure system design. This paper proposes a novel analysis method for resilience evaluation of cryptographic algorithms, which takes DES as example and reveals inherent algorithmic properties related to side-channel attacks. Collision and confusion coefficients are defined as the algorithmic parameters. The analysis shows that in addition to the side-channel leakage, another algorithm-dependent factor determines the effectiveness of side-channel attacks. With such factor considered, a metric is proposed to evaluate side-channel attacks and system resilience. Experiment results demonstrate the effectiveness and efficiency of the metric.