Resource Allocation Of Optical Network For Grid Application

Grid is an infrastructure for resource sharing. It comprises of numerous heterogeneous resourcethat are geographically distributed and form a super computer connected by Internet. With thedevelopment of grid computing, grid put forward higher requirements for the network, and theoptical network can be able to meet the gird requirements. So, the new technology of optical girdwas produced. Optical network resources as part of the optical grid, its large bandwidth, low delayand other features can effectively solve the problem of transmission grid task. At the same time, itcan greatly improve the processing speed of grid tasks. The processing of a job on a conventionalnetwork takes place on a single computer, whereas the processing of a job on a Grid may be splitbetween several resource nodes. So, grid resource allocation is an important problem, and that wewant to find a method about how to better and more economical allocation optical grid resources.Different strategies have been reported in the literature to discover and allocate resources thatare suitable for jobs to achieve performance up to user’s satisfaction. But, these papers rarelyconsider the economic principle of allocation resource problem. The thesis mainly researches arebased on the optical gird theory. And, In this thesis, the key problem of”resources allocation ofoptical network for grid application” is investigated in detail.The thesis introduced the development and current status of the optical gird, and discussedseveral key problems were existed in the resources allocation. For the economical problem ofresource allocation, in this thesis proposed two modified resource allocation algorithms that werebased on Vogel’s approximation method. First of all, an improved algorithm is proposed under thecondition of only considering the network resource allocation. By comparison with other algorithmsit was found that the improved algorithm of total computing cost results will be smaller. So the girdusers in the use of network resources cost is smaller. Second, the joint minimum cost algorithm isproposed under consideration of all grid resources. In this thesis, the effectiveness and economy ofthe method that was analyzed and verified by comparison with other algorithms, found that theimproved algorithm of economy will be better. The results show that these two kinds of algorithmscan achieve a rational allocation of resources, and allows users to get performance about the lowercost.