An Enhanced Adaptive Scoring Job Scheduling algorithm for minimizing job failure in heterogeneous grid network

Grid computing involves sharing data storage and coordinating network resources. The complexity of scheduling increases with heterogeneous nature of grid and is highly difficult to schedule effectively. The goal of grid job scheduling is to achieve high system performance and match the job to the appropriate available resource. Due to dynamic nature of grid, the traditional job scheduling algorithms First Come First Serve (FCFS) and First Come Last Serve (FCLS) does not adapt to the grid environment. In order to utilize the power of grid completely and to schedule jobs efficiently many existing algorithms have been implemented. However the existing algorithms does not consider the memory requirement of each cluster which is one of the main resource for scheduling data intensive jobs. Due to this the job failure rate is also very high. To provide a solution to that problem Enhanced Adaptive Scoring Job Scheduling algorithm is introduced. The jobs are identified whether it is data intensive or computational intensive and based on that the jobs are scheduled. The jobs are allocated by computing Job Score (JS) along with the memory requirement of each cluster. Due to the dynamic nature of grid environment, each time the status of the resources changes and each time the Job Score(JS) is computed and the jobs are allocated to the most appropriate resources. The proposed algorithm minimize job failure rate and makespan time is also reduced.