Performance and formal design: a process algebraic perspective

So far, most research in the area of formal methods has been focussed on the development of theories, methods, and tools for the design and analysis of functional, or qualitative, aspects of information-processing systems. Performance analysis, on the other hand, has always been concerned with the quantitative analysis of such systems. As a result each community has been doing its research mostly independently of the other, although a number of formal models of system behaviour have made their way into the world of performance analysis. First of all, there is the now established field of stochastic Petri nets and its application to performance modelling. More recently, there is a growing interest in the application of process algebraic techniques to performance modelling, and a number of proposals for timed, probabilistic, and stochastic process algebras have been put forward. At the same time we observe that the once clear distinction between the functional and performance properties of systems is getting blurred. With the technological means and the economical drive in place to offer a host of high-performance services to end-users there is a clear need to treat the quantitative quality-of-service parameters as requirements in functional specifications. The rapidly growing importance and proliferation of such systems not only implies a conceptual merging of functionality and performance, but also calls for the integration of qualitative and quantitative aspects in their design and realization. In our presentation we analyse to what extent the use of process algebraic techniques can contribute to an increased collaboration between the performance analysis and formal methods communities. Ideally, such collaboration should not only lead to further progress in performance analysis, but also contribute to the incorporation of performance analysis in an integrated discipline of formal design. Some of the more specific topics that we address are: design by transformation; derivation of performance models; compositionality; liveness and fairness; and true concurrency.