Paper 2022/174

How to Launch a Powerful Side-Channel Collision Attack?

Jiangshan Long, Changhai Ou, Yajun Ma, Yifan Fan, Hua Chen, and Shihui Zheng

Abstract

Benefiting from its independence of leakage model, side-channel collision attack is one of the most common distinguishers and attracts wide attention. Although several improvements have been given, its performance on attacking a single collision value has not been significantly improved. Its optimization and efficiency is still an open problem. To solve this, we theoretically analyze the quantitative relationship between encryptions and collisions in this paper, and propose an efficient side-channel attack named Collision-Paired Correlation Attack (CPCA) for low noise scenarios to guarantee that the side with fewer samples in a collision to be detected is completely paired. This optimizes the inefficient utilization of collision information in the existing collision attacks. Moreover, to further exploit the collision information, we maximize the collision pairing, and this optimization significantly improves CPCA and extends our CPCA to large noise scenarios. Finally, to reduce computation complexity, we further optimize our CPCA to a CPA-like distinguisher. Our further theoretical study fully illustrates that our CPCA provides the upper security bound of CECA, and experimental results fully show its superiority.