Cryptographic Randomness Testing of Block Ciphers: SAC Tests

Abstract

Block ciphers are designed to function as random mappings, making it essential for them to successfully pass statistical randomness tests. These tests evaluate whether the distribution of a test statistic, derived empirically through various data manipulations over states of the algorithm, aligns with the theoretical distribution for cryptographic randomness. Beyond this, evaluating the cryptographic properties of the algorithm is also important to ensure its security and reliability. One of the important cryptographic randomness properties is the Strict Avalanche Criterion (SAC), which assesses the impact of a one-bit alteration in the input over the output. In this work, we introduce new SAC-based tests to offer more reliable evaluation for the cryptographic randomness of block cipher algorithms. The tests are utilized for the application of AES, PRESENT, and CLEFIA block ciphers. The results are compared with Soto's evaluation methods, which are known for their comprehensive approach to block ciphers. According to this, it is apparent that our novel SAC tests improve upon Soto's results, thus providing a more comprehensive understanding of randomness.