Resource Management And Job Schedule In Campus Grid

With the prevalence of Internet, different kinds of equipment were connected in different manners by Internet to realize the share of the resources. Campus grid is the implementation of the grid technology in the campus area. It aims at providing unified and transparent service to the user by organizing all the computing resource of the campus in an efficient manner and hiding the internal heterogeneousness and dynamic. The whole campus grid is a three-tier system, which is composed of grid portal layer, grid middleware layer and computational resource layer. The grid portal layer acts as the grid resource broker and implements global scheduling. It provides a unified interface to the user and allows trusted users seamless access to computational resources and grid service by offering service such as user account management, security and authentication, job submission and schedule, job monitor and job result disposal. The grid middleware layer supports the basic grid service. In our system we use the grid security management, information service, fault detect service and resource allocation management in the Globus Toolkit which is the de facto standard for grid computing because of its wide acceptance and deployment worldwide. Computing resource includes all kinds of computing resource in the campus. Through Internet they were organized in different manners according to the type of the resource. In the grid system hierarchical structure was adopted in the resource management and in order to utilize grid resources effectively, grid job submission is completed by the resource brokers. The self-adapted grid schedule arithmetic will be adopted in this module. Under the premise of the job deadline, appropriate resource will be selected according to the computing time and computing ability of the resource, which guarantees the schedule veracity and use all the resource in an efficient way.From the result of the instance which has applied on the system the validity of the self-adaptable scheduling model can be demonstrated.