PUF-Based Random Number Generation

From security to randomized algorithms, there are many existing problems whose solutions are fundamentally based on the assumption that intrinsically pure random number sources exist. Pseudorandom number generators can imitate randomness sufficiently well for most applications, however, they still rely on some secret seed input. Hardware random number generators attempt to extract randomness directly from complex physical systems. In this way they create random outputs without requiring any seed inputs. In this paper we describe how to use Physical Random Functions (or Physical Unclonable Functions, PUFs) to create a candidate hardware random number generator. We present a short argument supporting the tenability of our methods and provide a brief evaluation of the “randomness” of numbers generated using PUFs through a series of statistical tests. These tests show promising possibilities for the use of PUFs in hardware random number generation.