As communication and computation become colossally embedded and interconnected - new demands on cyber-physical systems call for innovation in many areas including cryptography, privacy, security, spectral management, compression, sampling and actuation/sensing. Specifically, high data-rates and processor speeds challenge both the periodicity and unpredictability of pseudorandom number generators (PRNG). For some applications, minor, on-going tweaks to existing PRNGs are sufficient. However, as entropic demands increase, there is a need (in some cases) for ironclad, true random number generators (TRNG). If genuinely attainable at high data rates, these devices could permit provably secure communication, privacy and potentially optimal compression, sensing or sampling.

The fundamental burden associated with true randomness is that it is not provable. Although there are tests for evaluating the performance of PRNGs, any such test applied to TRNGs is categorically inappropriate. Engineering a source of provably true, random numbers conjures issues in philosophy, mathematics, computer science, statistics, thermodynamics, solid-state physics, quantum mechanics, information theory and dynamics. Although the philosophical proof that a finite set is truly random leads to conundrum or impossibility - there are widely accepted arguments based on first principles and best practices that are indicators that a TRNG is bona fide. This talk considers these first principle-based lines of reasoning to support claims of randomness. Additionally, a review of current literature and consideration of integrated circuit challenges is presented.