Analytical analysis of reliability for executing remote programs on idling workstations
C. Zhang, Cheng-Hsiung Yang · 1990
A discussion is presented of two control policies of workstation-based distributed systems that support fault-tolerant execution of remote jobs on idling workstations: checkpointing and noncheckpointing. An analytical analysis of the reliability and mean turnaround time of the execution of remote jobs is conducted for both control policies. In addition, the optimal time interval of checkpoints in the checkpointing policy is formulated on the basis of the given reliability and overhead of the system. Some observations on fault-tolerant features of each policy derived from the analysis are presented.>