Fault Tolerant Job Scheduling in Computational Grid

In large-scale grids, the probability of a failure is much greater than in traditional parallel systems [I]. Therefore, fault tolerance has become a crucial area in grid computing. In this paper, we address the problem of fault tolerance in term of resource failure. We devise a strategy for fault tolerant job scheduling in computational grid. Proposed strategy maintains history of the fault occurrence of resource in grid information service (GIS). Whenever a resource broker has job to schedule it uses the resource fault occurrence history information from GIS and depending on this information use different intensity of check pointing and replication while scheduling the job on resources which have different tendency towards fault. Using check pointing proposed scheme can make grid scheduling more reliable and efficient. Further, it increases the percentage of jobs executed within specified deadline and allotted budget, hence helping in making grid trustworthy. Through simulation we have evaluated the performance of the proposed strategy. The experimental results demonstrate that proposed strategy effectively schedule the grid jobs in fault tolerant way in spite of highly dynamic nature of grid