Economic-based Grid Resource Allocation Algorithm

Grid is a new computing platform after Internet and Web. Its main character is sharing resources. In this paper, we research and optimize the grid resource allocation schemes based on the important frame GRACE in economic grid. This objective of this research is to solve the problem of grid user negotiating price in grid resource allocation and to test the schemes proposed. Firstly, we introduce GRACE and models in resource allocation based on economic grid. Secondly, uniting the price fluctuation to reputation scheme, we propose one new resource allocation scheme and one distributed resource allocation scheme in economic grid. At last, we test these schemes, and show that these new methods can improve the resource service successful rate and the grid profits. The main contents are the resource price fluctuation algorithm, goal function in resource selection, and distributed resource allocation based on reputation and price fluctuation. The primary contributions of this thesis include:1. Grid resource allocation strategy based on price fluctuation and reputation method is proposed to solve the problem of GSP's price changed to adapt to the grid market. There are many influencing factors in grid Environment, such as the reputation, trade successful rate, the GSP's profit and so on. Once the trade is end, the price should be updated to adapt to the grid market. From the simulate test, we find this strategy can improve the trade successful rate and the grid's profit.2. Optimize the goal function. In economic grid, when GSC needs to select the most appropriate GSP, many factors may influent the GSC. So an optimal goal function is very important. In this paper, we propose a goal function for GSC, and the test results show that it is effective.3. We propose a distributed resource allocation strategy based on the virtual organizational domain (VOD). In this strategy, we considered the network load and distributed reputation based on VOD. The test results show that the strategy can decrease the network load and grid resource consumption.