Notes on non-Markovian Extension of Value Passing CCS

We define a non-Markovian extension of Value Passing CCS, based on PH-distributions. The approach consists in describing the execution time of each action with a PH-distribution, i.e. by associating a finite CTMC with a terminal, absorbing, state to each action of a process. Such CTMC characterises a random variable, namely the time needed to go from the initial state to the final one. It is well known that such distributions can be used for approximating, ideally, any general distribution and are closed under max. The choice of a CCS-like pattern of interaction is motivated by the fact that we are interested in service-oriented computing models and many process calculi for service-oriented computing are based on such a pattern of interaction.