On the Relation Between Reliability and Entropy in Physical Unclonable Functions

Physical Unclonable Functions (PUFs) are integral for generating unique signatures, secret keys, and device identification, leveraging inherent manufacturing process variability.Mathematically defined as functions linking inputs (challenges) to outputs (responses), PUFs exhibit random properties.Key properties for high-quality PUFs include intra-device entropy (random distribution of responses within the same circuit), interdevice entropy (random distribution across different circuits for identical challenges), and reliability (response consistency for identical challenges and the same circuit).Inter-device entropy and reliability may be influenced by design discrepancies, systematic variability, noise, and aging.This paper addresses the correlation between entropy and reliability, providing evidence from an extensive set of circuits featuring diverse Ring Oscillators supplied by Infineon.