An Efficient Resource Management and Scheduling Technique for Fault Tolerance in Grid Computing?

In Grid computing, scheduling techniques are based on queues, easy to implement but not very effective in mapping jobs after well determined parameters. If a certain user needs a specific resource, usually she/he has to log on directly on specific resource and to access the scheduling service. A more advanced method is to have a schedule-based approach. This approach is usually most effective when the submitted jobs are in batch mode (a new task will trigger either rescheduling, either will be placed in a temporary queue if it has a low priority, until a significant amount of jobs are queued, depending on the scheduling algorithm used). The proposed technique will give the user a single view where the user can log into the Grid resource management system, submit jobs and check the status of his jobs just like he does when he submits a job to any single resource.