The Research And Implementation Of Distribute Resource Collecting And Scheduling Based On Grid

Grid is a new generation of the Internet and a new kind of information technology facilities. At present many theories and technologies about Grid are not very mature, especially about the theories and technologies of Grid Resource Scheduling. As Grid Resource are distributed, heterogeneous, dynamic and autonomy, so how to improve the scalability of Resource Scheduling and improve the resource using ratio is the key point in Grid technology.This paper is based on the Gird Management Product EGO (Enterprise Grid Orchestrator) which is the product of Platform Computing Company, take Grid Computing and the technologies of Grid Resource Scheduling as background, and focus on the research of Enterprise Grid Resource Collecting and Scheduling technology.At the beginning, I studied the component of EGO Resource Collecting and Scheduling, analysis the current Collecting and Scheduling Policy, and then identified its limitations and shortcomings. Against the Collecting model limitations and shortcomings, I designed a new Collecting Model: Plug-in Collecting Model, which is based on the theory of polymorphism in object-oriented programming, which means that the Collecting Model was implemented in the format of Plug-in that can be loaded dynamically when the system is running, and this increased the scalability and flexibility of the Collecting Model for a Grid Resource Management System. And against the limitations and shortcomings of Scheduling Policy, I improved the Scheduling Policy, improve the resource utilization.Then I implemented the improvement design, and tested them through function testing. The test data proved that the new Collecting and Scheduling Policy can not only help improve the resource using ratio, but also improved the scalability and flexibility of the system, also this improvement did not affect the system's performance.