Publication

Abstract : Grid computing has emerged as the next generation parallel and distributed computing methodology that aggregates geographically distributed heterogeneous resources which have the potential to solve large scientific applications. In addition to the challenges of job scheduling, the effective utilization of resources and fault tolerance plays a major role due to the unreliable nature of the grid environment. Load balancing is a technique to enhance the resources, utilizing parallelism, exploiting throughput improvisation and to reduce the response time through an appropriate distribution of the application. Several existing algorithms address the grid system with no resource broker and few with single resource broker to provide solution with increase in efficiency with a variety of load variations. The proposed work holds a grid scenario with multiple resources brokers explicitly created for each set of grid resources interconnected with each other, categorized based on the capacities of the Processing Elements (PEs) to balance the load and to achieve reduced execution time of jobs that are being processed. A Weighed Rank Based Scheduling (WRBS) algorithm is proposed for the Meta broker to route the incoming jobs to sub- level of resource brokers and hence as a result an optimum solution is obtained thus achieving the effective scheduling and optimum load balancing with effective utilization of resources under heterogeneous grid environment. The simulation result shows significant results with a minimum of four percentage increase in efficiency through reducing the overall execution of jobs for minimum load and thereby shows an increase in efficiency for greater loads depending on the number of incoming jobs. © Research India Publications.