The Research Of Grid Resource Management And Scheduling Algorithm

In the past ten years, scientific computing has been shifting from centric todistribute and parallel pattern; recently this trend has extended to grid computing. Gridis an important development trend for parallel and distributed computing technologynowadays,aiming at realizing sharing the large quantity of heterogeneous resourceslargely distributed geographically. Because of the characters of distributed and shared,self-resembling, dynamic and diverse, autonomous and manifold management of agrid, resource management and scheduling under grid environments is one of the keyresearch fields of grid systems. Now the tools of grid simulation are gradually beingused to help the designers of the grid systems to validate the scheme and test theperformance of the grid system. Especially the simulation tools for the part ofresource scheduling have become the key point of the research in this field.In this paper, the basic concepts, characters and application fields of gird areintroduced;second, the popular grid infrastructures are described;then severaltypical grid technologies and resource management systems are discussed,focused on Nimrod-G grid resource management system:furthermore, a gird resourcetrading management algorithm is offered;also, a new price time optimizationalgorithm is presented based on the classical price optimization and time optimizationalgorithms. Because in the research of grid tasks scheduling, it is unnecessary to usereal systems to evaluate the performance and accuracy of these algorithms, usuallyemulators are used to handle with this kind of work. Several grid emulators arecompared in this paper, and GridSim—a grid modeling and emulation toolkit isintroduced in details.Thesis introduces the architecture of the grid simulator, themechanism and process of the modeling simulation and the realization method ofevery module in details. It includes the modules of user and resource, resourcescheduling police, grid environment, java code generation, code compilation andexecution and result analyses. The extensible grid simulation platform which theessay puts forward focusing on the part of grid resource scheduling, considers thetopology of the network, so the result is more agree with the real grid environment, and it sets up a good base for the future research in the field of grid simulation.Finally, with the help of GridSim toolkit, the emulation of grid resourcescheduling with QoS requirements is realized.The experiment results show that the new price and time optimization algorithmpresent in this paper is feasible. It improves the scheduling performance of schedulingabove the original price optimization and time optimization algorithms.