An automata-based approach for addressing erroneous behaviors and deadlocks in component-based systems

Abstract

In component-based modeling, a system is built by constructing several coordinating components. In turn, the communication principles among components are modeled by architectures that characterize the permissible interactions and the topology of the components. Several architectures impose also order restrictions to the appearance of the permissible interactions. The application of an architecture to a component-based system does not guarantee that the behaviors of the system meet all the requirements of the architecture. In turn, this can lead the system to actions that do not follow the intended order of the interactions, which we call erroneous behaviors. In this paper, we tackle the problem of such order violations by introducing a formal framework for building component-based systems that avoid erroneous behaviors. For this, we formalize the systems by nondeterministic finite automata and we model their architecture by a fragment of extended propositional interaction logic. This logic achieves to encode the order restrictions in the occurrence of the permissible interactions of the architecture. Given a component-based system with an architecture, we propose an automata-based method for constructing a respective system without erroneous behaviors. Our algorithm returns a system that is also free of deadlocks, meaning that the system does not reach states from which it can no longer progress.