Research On Resource Scheduling Policy In Environment Of Grid

Grid resource scheduling strategy is an important component of researching fields in the grid computation. It is effective for allocating resources to the matched grid jobs based on their characteristics. The scheduling results will be useful for optimizing the usage of grid systems. In addition, for users in the system, grid resource scheduling strategy will be configurably or transparently used to achieve goals that grid resources will be easily and effectively used.Therefore, resource scheduling strategy is the hotspot field in grid computation. Firstly, this paper introduces the basic conceptions and theories of grid system and scheduling system. Then, the used scheduling mechanism and criti- cal technology are addressed after GridSim, the grid resource scheduling simula- tion toolkits, is represented.Secondly, since grid jobs and resources have their own characteristics, the attributes of themselves are typically appointed as deadline of grid jobs and bandwidth of resources for researching scheduling algorithms. Through analyzing the basic thought of classical scheduling algorithms, double QoS constraint algorithms are proposed. Considering the performances and bandwidths of resources respectively, CF and LBF algorithms are proposed by analyzing the insufficiency of BF algorithm. Senior algorithm is proposed based on the forecast information of execution time for better scheduling results.Finally, algorithms addressed before are simulated by the GridSim toolkits. It is proved that senior algorithm is superior in solving deadline and bandwidth constraint problems by comparing and analyzing result data. Conclusions of research are summarized and existed issues are pointed out for future research.