Hybrid Resource Allocation Method for Grid Computing

The resource management system is the ´brain´ of a Grid. It manages the shared resource of the Grid and maps user jobs to the resources. Grid resource allocation is one of the critical functions affecting the performance of a Grid, because the number of jobs and amount of required resources are massive and quick responses to users are necessary in a real Grid computing environment. Many methods have been developed for Grid resource allocation. In this paper, we propose the Hybrid Resource Allocation method, based on the Least Cost Method (LCM) and Divisible Load Theory (DLT) method. The Hybrid Resource Allocation method is an improved form of the DLT method. We have developed a modified assignment strategy of DLT by integrating it with LCM. In this paper we describe and evaluate Hybrid Resource Allocation method for Grid resource allocation scenario. This paper also proposes a Grid scheduling model and includes a comparative performance analysis of our proposed Hybrid Resource Allocation method with the existing ones.