Comprehensive Method for Measuring Randomness in Pseudorandom Generators

Abstract

A novel method  is presented  through which a randomness index is used to measure the quality of the sequences produced by any pseudo-random number generator. To obtain this index, several statistics are combined, some known and others proposed in this research. The method is comprehensive in the sense that, to calculate the randomness index, large sets of data sequences delivered by the generator are considered as a whole. Millions of tests were carried out to generate sequences of pseudo-random numbers and the randomness  index was calculated for blocks of these sequences. The results obtained show that the proposed  procedure to measure randomness  is robust and far outperforms the most popular procedures, which makes it ideal for evaluating the behavior of high-performance generators such as those used in cryptography tasks.