An Adaptive Filter Mechanism of Random Number Generator in a Crypto Module

Abstract

Yet, when using a real random number generator (RNG) with only a hardware component, as required for statistical randomness, it is quite difficult to create an unbiased and stable random bit stream. Although the hardware generating processor generates an output bit stream quickly, if the software filter algorithm is not efficient, the RNG consumed a relatively large amount of time. This factor becomes the limiting condition when the RNG is applied. Accordingly, this paper proposes an adaptive filter approach to ensure the model of software filtering in the RNG processor of a crypto module. Therefore, an adaptive model is proposed that applies a filter algorithm, thereby consuming less time than the conventional filter algorithm scheme.