An algebraic view of interleaving and distributed operational semantics for CCS

Abstract

In this paper we describe CCS models in terms of categories of structured transition systems: we define two categories CCS and CMonRCCS for representing interleaving and “truly concurrent”, distributed aspects of CCS (without recursion). Among the objects of CCS and CMonRCCS we choose two standard models, called M and M respectively. We show that our interleaving model M essentially coincides with the classical transition system of CCS, while the distributed model M faithfully expresses the issues about decentralized control and multiple representations of agents discussed in a recent paper by the first author in collaboration with P. Degano and R. De Nicola [4]. Consistency of distributed and interleaving semantics is also proved.

The advantage of defining categories of models instead of simply models is that within the same category we can distinguish initial transition systems, whose transitions are proofs, and interpreted transition systems, which embody synchronization algebras. Another advantage of categories is the use of free adjoints of forgetful functors for constructing the models and for comparing transition systems having similar, but different, structures.

Research performed in part while visiting ESLAI with the support of the Italian Foreign Ministry, Programma di Cooperazione e Sviluppo.