Runtime Estimation Aware Scheduling Algorithm for Handling Deadline Based Jobs in Grid Environment

One of the major issues in Grid Environment is scheduling the deadline based jobs in the meta-scheduler level. In many approaches Bulk Scheduling is done in meta-scheduler by evenly distributing all the jobs present in the request handler queue to the available resources. All the submitted jobs will fall into the resource queue and get executed in the First Come First Served (FCFS) or Shortest Job First (SJF) fashion and it will works better for non-deadline based jobs where prioritization of job is not required. This type of scheduling algorithm will leads to starvation and increases the waiting time of deadline based jobs, and finally results in user dissatisfaction. Moreover, these types of scheduling algorithm in the meta-scheduler will not provide the feasible solution to the deadline based job. So a Runtime Estimation Aware Scheduling is proposed in the meta-scheduler, which provides high priority for deadline based jobs to dispatch first, during the scheduling process. Most algorithm will works good if the job have the runtime input and may not good enough if the job doesn’t have the runtime input. Our scheduling algorithm will works well with irrespective of jobs nature (either job with runtime or without runtime), obviously reduces the waiting time of deadline based (high priority) jobs and lights up the grid providers to achieve Service Level Agreement (SLA).