A Framework for Formal Reasoning about Open Distributed Systems

We present a framework for formal reasoning about the behaviour of distributed programs implementing open distributed systems (ODSs). The framework is based on the following key ingredients: a specification language based on the µ-calculus, a hierarchical transitional semantics of the implementation language used, a judgment format allowing parametrised behavioural assertions, and a proof system for proving validity of such assertions which includes proof rules for property decomposition. This setting provides the expressive power for behavioural reasoning required by the complex open and dynamic nature of ODSs. The utility of the approach is illustrated on a prototypical ODS.