Asymptotic Competitive Analysis of Task Scheduling Algorithms on a Fault-Prone Machine

Abstract

Reliable task execution in systems with machines that are prone to unpredictable crashes and restarts is challenging and of high importance. However, not much work exists on the worst case analysis of such systems. In this work, we analyze the fault-tolerant properties of four popular scheduling algorithms, Longest In System, Shortest In System, Largest Processing Time and Shortest Processing Time, under worst case scenarios on a fault-prone machine. We define three metrics for the evaluation and comparison of their competitive performance in the long run, namely, completed cost, pending cost and latency. We also investigate the effect of resource augmentation by increasing the speed of the machine. Finally, we compare their behavior for different speed intervals concluding that each of them behaves better than the rest under different circumstances and none is clearly the best.