Alternative of Strict Avalanche Criterion Test on Block Cipher Using Fisher’s Method

Abstract

Block cipher was one of the cryptography techniques used to provide information security, especially confidentiality. Strict avalanche criterion (SAC) was a desirable characteristic of block cipher. If a block cipher satisfies strict avalanche criterion, then each output bit should change with a probability of one-half whenever a single input bit was complemented. Avalanche matrix was constructed to ascertain whether a block cipher satisfied SAC. A block cipher satisfies strict avalanche criterion if each entry of avalanche matrix was equal to one-half. Dawson proposed statistically based methods for determining whether a block cipher satisfies SAC. In this paper, we proposed an alternative statistical test to test the hypothesis that a block cipher satisfies SAC. The approach used to develop the alternative test statistic was by testing equality of each entry of the mean of avalanche matrices to one-half and then combined using Fisher’s method. That statistic was minus two times sum of the logarithm of p-value, where the p-value was tail area probabilities of z-statistic from each entry of the mean of avalanche matrices. The simulation showed that the power of the alternative SAC test was higher than the Dawson SAC test.