Compatibility in a multi-component environment

Abstract

A distributed environment where many components interact may be functioning in a suboptimal manner due to two main factors: message loss and deadlocks. Message loss occurs when a component is not ready to receive as input a message sent to it. In the case of a deadlock, a system is indefinitely waiting for a message that never arrives. In Carmona and Cortadella (2002) [12] a theory has been presented for characterizing when a pair of systems is compatible in the sense that they can engage in a dialog free from these two problems. The theory developed was restricted to only two components, a particular mode of synchronization and a closed environment. In this paper we lift all these assumptions to define a general notion of compatibility in a multi-component environment. For the extended definition of compatibility, we use team automata as a modeling formalism which allows arbitrary synchronization strategies and iterative/hierarchical composition. Moreover, it is shown how the general definition of compatibility presented in this paper can be used to determine the compatibility problems that arise in a team automaton built on the basis of an arbitrary synchronization strategy.