Research On Grid Resource Management And Scheduling Based On Agent Technology

The Grid is a new network computation platform after World Wide Web, its aim is providing infrastructure of comprehensive sharing various resources for users. Comparing with traditional distributed system's resources, the resource of the grid has the characteristic of distribution, dynamic, heterogeneity and etc, so the grid resource management and distribution are complication. This thesis puts forward an agent-based grid resource management and negotiation scheduling model. In this model, resource is not distributed by system directly. Through the price mechanism, the model lets the resource provider negotiating price with the resource demander. If negotiation is success, the resource will be distributed. In the model, resource is looked as commodity. The users need adding their preference when they buy commodity. During the negotiation, the user's preference and the resource's proportion of supply and demand will affect the last price. This model applies market economy theory, attracting the resource owners adding into the grid and helping to realization a big grid. Furthermore, the model imports mobile agent into resource management and negotiation process, adapting the grid's characteristic of distribution, dynamic, heterogeneity.