Performance analysis of a token-based distributed mutual exclusion protocol

The authors develop a methodology by which an approximation to a polling system model is used to analyze the performance of a simple circulating token mutual exclusion protocol. This protocol is chosen for a first attempt to apply performance measurement to mutual exclusion algorithms primarily because of its simplicity. A simple queueing model for a distributed protocol with a circulating token is presented. The analysis is tractable for arbitrarily large networks, and the results produced seem to be reasonably accurate.<
>