Grid Resource Management And Allocation Research Based On Bargaining Mechanism

Grid is an important development trend for parallel and distributed computing technology nowadays, aiming at realizing sharing the large quantity of heterogeneous resources largely distributed geographically. Resource management is one of the important techniques for grid. However due to heterogeneous, distributional and dynamic characteristics of grid, traditional resource management method is not applicable in grid environment. In addition, to attract resource consumers and providers of various kinds to join grid, their separate needs must be satisfied. At the same time, a great deal of economic theory studies have proved that, under many circumstances of little central guidance or without central guidance, market mechanism could be able to acquire the optimized resource allocation results by using economic methods to manage grid resource. This paper analyzes briefly the relevant research on applying economic models to allocate the resource in distributed system of recent few years. The paper continues with the introduction of bargaining theory and study on a bargaining negotiation model with incomplete information based on time restriction.Bargaining negotiation algorithm has been presented and analyzed for its performance through testing. Founded on this, the paper putsforward a grid resource management model based on bargaining mechanism that controls the resource by using the most basic bargaining method in economics. Grid consumers and resource could constitute bargaining policies for mutual negotiation according to their own needs. The task is completed finally that user agent selects the resource that most satisfies users' requirements. Compared to other resource management system based on economic model, the dominant advantage of this model lies in its simplicity and agility that it not only reflects the wills of resource provider, but also meets the requirements of users.