Quality Of Service Integrated Resource Selection Framework For Computational Grid

Abstract

The Computational Grid enables global sharing of computing newlineresources and provides inexpensive access to high end computational newlinecapabilities The integration of heterogeneous architectures should deliver newlinehigher levels of quality of service for the commercial computing applications newlineto benefit from the grid technology The present grid environment is oriented newlinetowards besteffort service but commercial environments demands guaranteed newlineand reliable services with committed service level agreements newlineResource selection is one of the most challenging problems in a newlinedynamic and heterogeneous environment Most of the resource selection newlinealgorithms select resources from the resource pool based on the rank of the newlineresources In computational grids the grid users are concerned with achieving newlinevarious qualities of service such as response time throughput availability newlinesecurity time and cost limits Provisioning of suitable resources to satisfy the newlinecompeting and conflicting needs of both the resource provider and resource newlineconsumer is the core goal of grid resource management and is the focus of newlinethis research work newlineIn this research work a Quality of Service Integrated Resource newlineSelection Framework is developed to deliver guaranteed services to grid newlineclients and avoid runtime failures job migrations and SLA renegotiations newlineDifferent components of the resource selection framework are evaluated using newlinedifferent mechanisms and integrated to provide reliable service to grid clients newlineThe proposed solutions are tested in the simulated grid environment using newlineGridSim toolkit 40 The job success rate resource utilization and job newlinecompletion time are the performance metrics evaluated for the proposed newlinestrategies and are compared with the existing MinMin strategy and GridWay newlinemetascheduler newlineThe Guaranteed Service Resource Selection GSRS Framework newlinedeveloped in this work selects appropriate computational and network newlineresources to successfully complete the client job within the user stated time newlineand cost limits dynamic SLA is presented by newlineexchange of templates and clie