Towards Quantifying Information Leakage of Timing Attacks in Cyber-Physical Systems

Timing side-channel attacks exploit variations in execution time to infer confidential information from a system’s behavior. In this paper, we develop a formal framework for analyzing timing side-channel leakage in cyber-physical systems by modeling their digital twins. We rely on a process-algebraic calculus for digital twins to express timing-sensitive behaviors and define its semantics using probabilistic Markov models. This enables rigorous quantification of information leakage based on execution timing variations. By leveraging entropy-based analysis, we measure an attacker’s ability to infer secret states from observable execution delays.