Stochastic φ-calculus

Abstract

We extend the φ-calcuhis, a model of concurrent processes based on the notion of naming, to cope with performance modelling. The new language is called the stochastic φ-calculus (Sk). We obtain a more expressive language than classical stochastic process algebras because Sir allows one to describe dynamically reconfig-urable or mobile networks. The semantics of Sir is given in the classical structural operational approach. In order to effectively compute performance measures, we define a stratified transition system that is finitely branching. We give a transition rule to directly yield a continuous time Markov chain from an Sir specification, with no transition system manipulation.