Design and Testing of RNG

Random Number Generators (RNGs), are essential components for every cryptographic system and application. In this work I will focus my attention on the best feasible approximation of an ideal RNG, defined as True Random Number Generator (TRNG). This module is the basic primitive to build of every crypto-system. It is mandatory also in the Pseudo Random (Algorithmic) world. Their realization exploits specific non-deterministic physical phenomena from various branch of physics. I will show a strong mathematical model, from which I will deduce a collection of statistical tools, like Fourier transformation, Entropy and Hamming distance. I will explore “ad hoc” methodologies for design RNG: analog and digital solutions. I will explore this filed and present the theoretical foundations, defined as Statistical Hypothesis Testing (SHT) and I will introduce some test suite: NIST SP800-22, DIEHARD, dieharder. Finally I will propose the core of this work: the design and the implementation of a CMOS on-chip test module.