Distributed optimized resource reservation scheme for grid computing

Abstract

The idea of grid computing originated in the scientific community and was initially motivated by processing power and storage intensive applications. The basic objective of grid computing is to support resource sharing among individuals and institutions within a networked infrastructure. Managing various resources in highly dynamic grid environments is a complex and challenging problem. There are works for resource management in different areas of computer science. Some approach uses algorithms for resource management to apply in grid but fails to provide any generalized solutions for grid environment. Most of the approaches are based on a simple architecture considering computer as the main resource in their system. But the real architecture of grid computing is a complex one by considering various resources of any computer during resource management. In grids sometimes assurance is needed for successful completion of jobs on shared resources. Such guarantees can only be provided by reserving resources in advance. So resource reservation is an integral part of resource management system for grid. Moreover the cost for providing resource as services will play a significant role in near future when resource sharing will be popular and inevitable but so far there is no work regarding cost optimization model for grid computing. In this thesis we provide a future reservation supported and cost optimized novel resource management system for grid environment considering its real complex architecture. We demonstrate our claims by conducting a detailed performance evaluation and comparing with the existing system on real workload traces for grid computing.