A Survey and analysis of algorithms for the detection of termination in a distributed system

This paper looks at algorithms for the detection of termination in a distributed system and analyzes them for effectiveness and efficiency. A survey is done of the published algorithms for distributed termination and each is evaluated. Both centralized distributed systems and fully distributed systems are reviewed. The algorithms are analyzed for the overhead and conclusions are made about the situations in which they can be used, i.e. an operating system, a real-time system, or a user application. An original algorithm is presented for the asynchronous case with first-in-first-out message ordering. It allows any process to initiate detection of termination and makes use of multiple tokens