Two improved multiple-differential collision attacks

In CHES 2008, Bogdanov proposed multiple-differential collision attacks which could be applied to the power analysis attacks on practical cryptographic systems. However, due to the effect of countermeasures on FPGA, there are some difficulties during the collision detection, such as local high noise and the lack of sampling points. In this paper, keypoints voting test is proposed for solving these problems, which can increase the success ratio from 35% to 95% on the example of one implementation. Furthermore, we improve the ternary voting test of Bogdanov, which can improve the experiment efficiency markedly. Our experiments show that the number of power traces required in our attack is only a quarter of the requirement of traditional attack. Finally, some alternative countermeasures against our attacks are discussed.