PEPA - Performance Evaluation Process Algebra

Abstract

The probabilistic and stochastic models we have presented in previous chapters represent system behaviour but not its structure, i.e., they take a monolithic view and do not make explicit how the system is composed and what are the interacting components of which it is made. In this last chapter we introduce a language, called PEPA (Performance Evaluation Process Algebra), for composing stochastic processes and carrying out their quantitative analysis. PEPA builds on CSP (Communicating Sequential Processes), a process algebra similar to CCS but with slightly different primitives. In particular, it relies on multiway communication instead of binary (I/O) communication. PEPA actions are labelled with rates and without much effort a CTMC can be derived from the LTS of a PEPA process to evaluate quantitative properties of the modelled system. The advantage is that the PEPA description of the CTMC remains as a blueprint of the system and allows direct re-use of processes.