Reliability-aware resource management for computational grid/cluster environments

The collective resource utilization achieved through grid computing is critical to the overall computing capacity of the collaborative community and should be guaranteed. Especially, in an existing environment where job sites are Beowulf cluster systems, a service node failure may render the whole system outage. Current grid fault tolerance techniques only address these issues in an opportunistic fashion. Thus, there is a need for complementing these approaches by pro actively handling failures at a job-site level, ensuring the system high availability with no loss of user submitted jobs. Our grid-aware cluster resource management effort was motivated by the fact that a cluster turns into a popular job site in the computational grid environment. We propose a solution dealing with fault tolerance at the service level complementing the task-based solutions as being done in some recent studies. We discuss various service availability issues related to the grid, and preliminary results obtained while implementing the smart failover and transparent job-queue replication mechanism and the automated grid installation package. Our report entails the benefits outweighing acceptable overhead after implementing our proof-of-concept framework.