Software Engineering Institute

We are concerned with reasoning about the behavior of assemblies of components from models of components and their patterns of interaction. Behavioral models of assemblies are "composed" from behavioral models of components; the (composed) assembly model is then mapped to one or more reasoning frameworks, each suitable for a particular kind of analysis. Information relevant to each reasoning framework must be faithfully rendered in the assembly model, and it must be preserved under interpretation to the reasoning framework. Our concern in this paper is the faithful rendering of concurrency. Our approach makes use of information provided by components and extracted from static assembly topologies to faithfully model real concurrency. The result is more effective analysis.