SAAC - Kriptosistemlerin analizi için istatistiksel mutlak çığ-etkisi kriter testi

Abstract

Cryptosystems using binary radix (bit) arrays must first meet the strict avalanche criterion, which prevents correlation between input and output values in order to resist differential attacks. For this criterion, which is also called confusion/diffusion and butterfly-effect in the common literature, it is expected that two inputs with only one-bit difference from each other will cause half the total number of bits to differ in the output bits. While the literature includes the difference of at least half of the output bits, it does not include that changing all the bits will cause the binary complement and thus another undesirable result. The Statistical Absolute Avalanche Criterion (SAAC) test proposed in this study allows the statistical determination of a lower limit and the establishment of hypotheses. The SAAC test is a tool that can be easily used in the analysis of cryptosystems, especially encryption methods and message hash functions.