Replication Based Job Scheduling in Grids with Security Assurance

Security assurance is a critical requirement for QoS or SLA satisfactions in risky grid environments because jobs may be scheduled to multiple machines across different distributed administrative domains. Unlike conventional methods using fixed-number job replications, in this paper, we propose a security-aware parallel and independent job scheduling algorithm based on adaptive job replications to make sure the job scheduling decision secure, reliable and fault tolerant. In risky and failureprone grids, the replication number is changed according to the current security conditions and the end-user settings. Simulation results show that it is robust due to its adaptive job replications and rescheduling mechanism. The performance results also show that it is better than nonsecurity-aware scheduling algorithms based on fixednumber job replications.