A Framework for Achieving Better Load Balancing and Job Scheduling in Grid Environment

Traditional multiprocessors Super computer systems do have certain constraints viz computing elements, storage space etc. These constraints can be minimized in Grid environment provided a proper job scheduling and load balancing algorithm is available. The existing job scheduling and load balancing are either in fully centralized or fully decentralized in nature. In this paper, we propose a scheme that focuses on both Local Area Grid and Wide Area Grid efficiently. If Local Area Grid is overloaded then only jobs are transferred to wide Area Grid and hence a well job monitoring is carried out to achieve better parallelism in Grid Environment that saves valuable clock cycles needed for job execution. The proposed algorithm has been studied by conducting simulation using Remote Method Invocation (RMI) paradigm.