RELIABILITY-ANALYSIS FOR THE EXECUTION OF REMOTE JOBS IN A WORKSTATION-BASED ENVIRONMENT

Many workstation-based distributed systems allow programs to be execut ed on remote machines for effective utilization of system resources. U sually, the control policies in these systems force a remote job be di scontinued by the arrival of local jobs to guarantee the autonomy of e ach workstation. Therefore, one special concern in the design of such systems is tile fault-tolerant aspects for the execution of remote job s. In this paper, we discuss two control policies of workstation-based distributed systems, the checkpointing and non-checkpointing policy, which support fault-tolerant execution of remote jobs on idling workst ations. An analytical analysis of the reliability and mean turnaround time of the execution of remote jobs are conducted for both control po licies. The optimal time interval between checkpoints in the checkpoin ting policy is formulated based on the given reliability and overhead of the system. In addition, several sample results derived from these analyses are compared with the outcome of corresponding simulation pro grams. Some observations of fault-tolerant features of each control po licy are then presented as guidelines for the future development of su ch workstation-based distributed systems.