In this paper we describe two differential fault attack techniques against Advanced Encryption Standard (AES). We propose two models for fault occurrence; we could find all 128 bits of key using one of them and only 6 faulty ciphertexts. We need approximately 1500 faulty ciphertexts to discover the key with the other fault model. Union of these models covers all faults that can occur in the 9th round of encryption algorithm of AES-128 cryptosystem, One of main advantage of proposed fault models is that any fault in the AES encryption from start (AddRoundKey with the main key before the first round) to MixColumns function of 9th round can be modeled with one of our fault models. These models cover all states, so generated differences caused by diverse plaintexts or ciphertexts can be supposed as faults and modeled with our models. It establishes a novel technique to cryptanalysis AES without side channel information. The major difference between these methods and previous ones is on the assumption of fault models. Our proposed fault models use very common and general assumption for locations and values of occurred faults. © International Association for Cryptologic Research 2006.
Mendeley helps you to discover research relevant for your work.
CITATION STYLE
Moradi, A., Shalmani, M. T. M., & Salmasizadeh, M. (2006). A generalized method of differential fault attack against AES cryptosystem. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4249 LNCS, pp. 91–100). Springer Verlag. https://doi.org/10.1007/11894063_8