Survey of true random number generation circuits

In the field of computing, random numbers in the form of binary strings are generated and are used as keys to mask or encode data in a way that it can be decoded only by authorized parties using the same key. Such end-to-end encoding or encryption can be used in various fields such as mobile communications, e-banking, email access, internet trade etc. and is known as Random Number Generation. A Hardware or True Random Number Generator (TRNG) is a circuit or a device that generates random numbers from a hardware physical process; unlike Pseudorandom Number Generators (PRNGs) that generate them purely through computer programs. A TRNG is much more resistant to cryptographic attacks that try to get to the core of random number generation process and explore exactly how the numbers are generated thus proving that the generator is not truly random and secure; but unlike PRNGs, TRNGs are almost always unpredictable and almost completely resistant to these attacks. The aim of this thesis is to explore some of these TRNG circuits and compare them with respect to technique of random number generation, implementation and some reliability performance metrics. The emphasis is on circuits that use intrinsic, in-circuit physical non-deterministic states and events as source for random number generation, for e.g. those implementable on CMOS hardware like digital/ analog/ hybrid ASICs or FPGAs.