Analysis of a time-out protocol and its applications in a single server environment

The use of time-out protocols can improve distributed system performance by utilizing equipment and resource-time that otherwise would be wasted. In time-out protocols for traditional computer communication networks, set-up time reduction is a major issue. An interesting area of development in networked production and test systems is the inclusion of costs other than set-up time in the evaluation, such as scrap rates and rework, or the effect on neighbouring or upstream client work-centres. The applications of time-out protocols in single, shared server production environments, such as automatic guided vehicle systems (AGVs) and shared testers in assembly processes, are described in this article. Analytical models have provided some insight into the application of time-out protocols but are cumbersome. The focus of this investigation has been in the context of networked task administration, including: (1) quantitative guidelines to evaluate time-out protocols in the context of task administration protocols, and (2) a new model for numerical evaluation, identifying the recommended time-out threshold and the regions where time-out protocols are, or are not, useful. Discrete event simulation models and empirical data from recent research on task administration protocols are described.