Study Of Grid Resource Scheduling Based On Model Of Petri Net

Grid computing is the potential of the next generation of platform,which will be used in field of science computing, engineering computing and some other large-scale computing. Grid core service is an important component of the grid computing. As a centrum of entire grid computing, grid core service connects grid understratum groupware to top groupware and coordinate entire grid system to work effectively, so it is very important and useful to research grid core service technology. Grid's resource scheduling technology is a part of grid core service technology. A good strategy of grid resource scheduling,which is in order to get the best performance and reduce total running time and total expense in grid computing, is needed to allocate grid resource. However, as the characters of grid resource are distributed and shared, self-resembling, dynamic and diverse, autonomous and manifold and so on, resource management and scheduling under environments is one of the key research fields of grid system.Petri net is a kind of good tool to describe concurrent and asynchronous system, there are many methods to analyze Petri net, such as reachability graph, reachability tree, incidence matrix and so on. However, its correctness can be proved by mathematics method. Compared with the characters of gird, it is convenience to use Petri net to describe grid.The paper mostly expatiate the preparation, design process and testing about the resource scheduling method which based on improved Min-Min algorithm. The major work is discussed as below:The basic concept, characters and architecture of grid computing are introduced; then the basic theoretic of Petri net is introduced, where workflow technology is expatiated on focus.A model of Grid Computing is set up,where using Extended High Level Time Petri Net. Then an improved model is put forward, which divides the tasks into two parts according to quality of services. The improved model is used in improved Min-Min algorithm, which could simplify the algorithm and also improve load.The algorithm is simulated by GridSim, and the factor on scheduling algorithm is analyzed. The experiment results show that the modified algorithm in this paper is feasible. It could improve the scheduling performance in a certain extent.