Result error detection on heterogeneous and volatile resources via intermediate checkpointing

Abstract

Desktop grids use the free resources in Intranet and Internet environments for large-scale computation and storage. While desktop grids offer tremendous computational power and a high return on investment, one critical issue is the validation of results returned by participating hosts that are volatile, anonymous, and potentially malicious. We conduct a benefit analysis of a mechanism for result validation that we proposed recently for the detection of errors in long-running applications. This mechanism is based on using the digest of intermediate checkpoints, and we show in theory and simulation that the relative benefit of this method compared to the state-of-the-art is as high as 45%.