The Study, Based On The Grid Economic Model Of Resource Allocation Strategies

Grid computing is the next generation and potential computing platform for solving large scale computing like scientific computing, engineering computing and business computing. By Grid technology, people can combine with servers, store systems and network. Then they can be made up of a super computing system and provide high quality services of computing. However, Grid resources are belonging to different providers, distributed in different places, heterogeneous and changeable. Therefore, it is not easy for the resource distribution in Grid environments. Especially, after combining Grid with economy, assigning the appropriate resource to tasks will definitely become a complex issue because Grid services consumers have different targets, deadlines, priorities, and budgets. Then, the conventional resource distribution methods can hardly catch up with the development of Grid economy in this complex Grid environment. Therefore, the research of resource distribution in Grid economy environments is becoming a research attraction in recent years.This paper firstly does some study and analysis on Grid economy architectures GRACE (Grid Architecture for Computational Economy). According to the existing architectures' characters, the RPFM (Resource Pricing Fluctuation Module) is designed based on GRACE. In addition, this paper proposes a resource distribution strategy added the mechanism of resource pricing fluctuation. Then, based on the research of conventional reputation control models, this paper gives a new reputation evaluation model which is adapted to the Grid economy environment and introduces the implementation of the RCM (Reputation Control Module) which is based on the new model. At last, this paper proposes a distributed reputation control architecture which spans several VOD (Virtual Organizational Domain) and designs a resource distribution policy based on the reputation.Through simulation experiments in GridSim environment, the results show that the improvements of Grid economy architecture and resource distribution strategies would improve the both benefits between GSP (Grid Service Provider) and GSC (Grid Service Consumer).