Research On Task Scheduling Techniques Of Grid Computing Based On Economic Models

Grid Computing is a hot technique developed in recent years, which integrates worldwide Internet resource into a virtual supercomputer, accomplish compute resource, storage resource, data resource, information resource, and knowledge resource to full share in global. Job scheduling is a hard problem in Grid System. A good job scheduling policy can reduce the executing time and the cost of jobs, improve QoS and improve the throughout of the whole Grid system. Using models in mature economic theory is benefit for Grid resource intelligent agent.This paper firstly introduce the conception of Grid, summarize the status quo and the development trend of Grid in recent years, and expatiate the importance of the job scheduling in the research of Grid Computing, then introduce the architecture GRACE, Nimrod-G system and the scheduling policy based on it. The three heuristic algorithms in the DBC Scheduling Policy can each optimize one target of the two QoS targets (time optimization or cost optimization).At last we simulate our algorithm with the aid of Gridsim toolkit and show the different effort of the algorithms. And simulate static and dynamic scheduling algorithm, then compare with each other vertically and horizontally to find the shortages and advantages of the algorithm, prove the economic model's regulatory function.