Quantifying Impacts of Resource Heterogeneity on Grid Performance

Abstract

Grid computing has evolved into an important discipline within the computer industry by differentiating itself from distributed computing through an increased focus on resource sharing, co-ordination, manageability and high performance. Grid computing combines open, shared, geographically distributed and heterogeneous resources to achieve high computational performance. The objective of the grid computing is to solve large problems which can not be solved by single CPU by achieving high computing performance by optimal use of geographically distributed heterogeneous idle resources. These resources may belong to different institutions, different domains, have different usage policies and pose different requirements on acceptable requests. However, the major challenges in such highly heterogeneous and complex computing environment are to design an efficient resource allocation and management infrastructure. So resource allocation strategies for such system should be smart, efficient, robust, and scalable. Resources are heterogeneous due to differences in hardware components, differences in grid software environments, and different administration have different policies for sharing of resources. Therefore, resource heterogeneity, dynamic load on resources, task runtime prediction uncertainty, task-to-resource ratio and resource sharing in the grid environment affects application performance. In this thesis, we have focused on grid environments typically populated with large number of heterogeneous resources. We have investigated and quantified the impacts of resource heterogeneity by executing grid application on homogeneous and heterogeneous resources. Results of our analysis show; when we increase resource heterogeneity in grid environment, the performance of the grid environment decreases as compared to homogeneous environment.