Continuously analyzing finite, message-driven, time-synchronous component & connector systems during architecture evolution

Abstract

Understanding the semantic differences of continuously evolving system architectures by semantic analyses facilitates engineers during evolution analysis in understanding the impact of the syntactical changes between two architecture versions. To enable effective semantic differencing usable in practice, this requires means to fully automatically check whether one version of a system admits behaviors that are not possible in another version. Previous work produced very general system models for message-driven time-synchronous (MDTS) systems that impede fully automated semantic differencing but very adequately describe such systems from a black-box viewpoint abstracting from hidden internal component behavior. This paper presents a system model for MDTS systems from a white-box viewpoint (assuming component implementation availability) and presents a sound and complete method for semantic differencing of finite MDTS system architectures. This method relies on representing (sub-)architectures as channel automata and a reduction from the semantic differencing problem for such automata to the language inclusion problem for Büchi automata. The system model perfectly captures the logical basics of MDTS systems from a white-box viewpoint and the method enables to fully automatically calculate semantic differences between two finite MDTS systems on push-button basis, yields witnesses, and ultimately facilitates semantic evolution analysis of such systems.