Group-based scheduling scheme for result checking in global computing systems

Abstract

This paper considers the problem of correctness to fault-tolerance in global computing systems. Global computing system has been shown to be exposed to intentional attacks, where authentication is not relevant, network security techniques are insufficient. To guarantee correctness for computation, fault-tolerance schemes have been used to majority voting and spot-checking but these schemes intend to high computation delay because are not applied scheduling scheme for result checking. In this paper, we propose a new technique called GBSS(Group-Based Scheduling Scheme) which guarantee correctness and reduce computation delay by using fault-tolerant scheduling scheme to the result checking. Additionally, simulation results show increased usable rate of the CPU and increased performance of the system. © Springer-Verlag Berlin Heidelberg 2005.