Improved Leakage Model Based on Genetic Algorithm

The classical leakage model usually exploits the power of one single S-box, which is called divide and conquer. Taking DES algorithm for example, the attack on each S-box needs to search the key space of 2 in a brute force way. Besides, 48-bit round key is limited to the result correctness of each single S-box. In this paper, we put forward a new leakage model based on the power consumption of multi S-box. The implementation of this method is combined with genetic algorithm. In DES algorithm, we can establish leakage model based on the Hamming distance of summing up 8 S-boxes. The genetic algorithm can search the key space of 2 to complete the attack of 8 S-boxes at the same time intelligently. And we also experimentally validate the fact that the leakage model of 8 S-boxes can decrease about 60% number of traces which is needed in the classical based on one single S-box in time domain and it also decreases about 33% number of traces in frequency domain. The IC card which is used in experiment is the training card 8 provided by Riscure Company.