Efficient Hybrid Random Number Generator Based on Keccak

Abstract

Random numbers are used in many areas such as cryptographic applications, statistics and game theory. These numbers are the most important part of these applications. There is a need for secure random numbers to guarantee the safety of these systems. In this study, a hybrid random number generator was designed to produce real random numbers that are safe and efficient. The generator consists of a deterministic part and an additive input. The Keccak hash algorithm has been redesigned for the deterministic part of the generator. Keccak is set as the latest standard for Hash algorithms. Additional inputs were produced with an approach based on the ring oscillator. Additional inputs to this generator prevent the number from being reproduced and guessed. The generated numbers satisfy the security requirements for cryptographic applications. Moreover, these numbers were shown that did not contain statistical weaknesses by NIST and autocorrelation tests. In addition, the developed system is not lost data, but it is 2 times more efficient. As a result, a productive random number generator has been developed that is safe to use for many applications.