Encoding CSP into CCS (Extended Version)

TL;DR

This paper explores two different methods to encode CSP's multi-way synchronization into CCS's two-way synchronization, analyzing their properties and efficiency, with implications for process calculus translations.

Contribution

It introduces two distinct encodings of CSP into asynchronous CCS, comparing their structural and behavioral properties, including a centralized and a decentralized approach.

Findings

Both encodings satisfy Gorla's criteria except for compositionality.

The centralized encoding uses a top-level context and weak bisimilarity.

The decentralized encoding is more efficient and ensures coupled similarity.

Abstract

We study encodings from CSP into asynchronous CCS with name passing and matching, so in fact, the asynchronous pi-calculus. By doing so, we discuss two different ways to map the multi-way synchronisation mechanism of CSP into the two-way synchronisation mechanism of CCS. Both encodings satisfy the criteria of Gorla except for compositionality, as both use an additional top-level context. Following the work of Parrow and Sj\"odin, the first encoding uses a central coordinator and establishes a variant of weak bisimilarity between source terms and their translations. The second encoding is decentralised, and thus more efficient, but ensures only a form of coupled similarity between source terms and their translations.