Uniform similarity measures in Grid resource discovery

Resource discovery is one of the basic and key aspects in grid computing, which provides the suitable resources satisfying the requirement of userspsila applications. At first, this paper reviews some novel similarity measures and presents an uniform one to determine the similarity between two objects, especially two grid resources. The resource discovery of grid is just based on the proposed similarity measures. And then a formal model of grid resources with three layers is proposed for effectively and efficiently organizing and discovering all kinds of heterogeneous resources. This is helpful to reduce the cost of time and correspondence in discovering resources through describing resources, abstracting resources and providing unified system images. Thirdly, over the organization of resources, the scheme of resource discovery is also described. Finally, analyses show that this approach is more feasible than the pure distance-based ones.