Language-based opacity in modular discrete event systems: Compositional secret-based verification using labeled petri nets

Abstract

This work focuses on verifying language-based opacity within modular discrete-event systems. We consider a distributed system that is modeled as a composition of multiple interacting modules, each modeled by a labeled Petri net. Ensuring confidentiality in such systems is critical for cyber-physical systems and industrial networks, where unauthorized inference of sensitive data can lead to security breaches. We introduce a new definition of language-based opacity for modular systems and propose three secret-based verification methods that avoid the construction of the monolithic system through parallel composition. Our approach includes three methods: (1) global secret verification via observer synchronization; (2) local, module-level secret verification; and (3) an iterative composition optimization that avoids building the entire modular system, yielding significant computational savings. Experimental results on a benchmark smart manufacturing system demonstrate the practical efficiency of our approach, showing orders-of-magnitude improvement in verification time and memory usage over traditional monolithic approaches.