Font Size: a A A

The Comparison Between A Commodity Market Model And An Auction Model In The Computation Al Grid

Posted on:2006-05-16Degree:MasterType:Thesis
Country:ChinaCandidate:Y M SunFull Text:PDF
GTID:2168360155953125Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
Computational Grids, emerging as an infrastructure for next generation computing, enable the sharing, selection, and aggregation of geographically distributed heterogeneous resources for solving large-scale problems in science, engineering, and commerce. Although the Grid develops rapidly, the payoff in the management of resources and application scheduling is unsatisfied. In this paper, I investigate G-commerce —the problem of dynamic resource allocation on the Grid in terms of computational market economies in which applications must buy the resources they use from resource suppliers using an agreed-upon currency. Farming the resource allocation problem in economic terms is attractive for the following three reasons: first, the resource usage is not free. Second, the infrastructure of economic models and Grid are similar, and by formulating Grid resource usage in market terms, we are able to draw upon a large-body of analytical research from the field of economics and apply it to the understanding of emergent Grid behavior. Last, if resource owners are to be convinced to federate their resources to the Grid, they must be able to account for the relative costs and benefits to doing so. Since the management of resources and application scheduling based on economic models are superior to traditional methods, there are more and more researches on it. They use economic methods to manage geographically distributed heterogeneous resources. There are many existed economic models, such as the commodity market model, the posted price model, the bargaining model, the tendering contract-net model, the auction model, the bid-based proportional resource sharing model, the community/coalition/ bartering model, etc. We can develop different resource allocation methods based on each models, but which is the best model for Computational Grid system is still a problem. Auctions are easy to implement in a distributed setting, and resource allocation based on auctions is implemented with both single-auctioneer and multi-auctioneer. This thesis tells about the basic principles of each economic model, and filters out commodity market model and auction model considering computational grid infrastructure, management of resources and application scheduling. It simulates a Grid environment using GridSim by referring the real Grid to do some research on resource allocation and application scheduling based...
Keywords/Search Tags:Computation
PDF Full Text Request
Related items