On Entropy and Bit Patterns of Ring Oscillator Jitter

TL;DR

This paper analyzes the entropy and bit patterns of ring oscillator jitter, providing methods to evaluate randomness quality, including a new entropy bound, to improve physical randomness sources in cryptography.

Contribution

It introduces efficient algorithms for evaluating entropy and bit patterns of ring oscillator noise, along with a new lower bound estimation formula for entropy.

Findings

Algorithms outperform Monte Carlo in speed and accuracy

Proposed entropy bound applies broadly to ring oscillator sources

Evaluation methods support standards like NIST 800-90B and AIS-31

Abstract

Thermal jitter (phase noise) from a free-running ring oscillator is a common, easily implementable physical randomness source in True Random Number Generators (TRNGs). We show how to evaluate entropy, autocorrelation, and bit pattern distributions of ring oscillator noise sources, even with low jitter levels or some bias. Entropy justification is required in NIST 800-90B and AIS-31 testing and for applications such as the RISC-V entropy source extension. Our numerical evaluation algorithms outperform Monte Carlo simulations in speed and accuracy. We also propose a new lower bound estimation formula for the entropy of ring oscillator sources which applies more generally than previous ones.