Scalable Virtual Organization for Grid System using Group Similarity Function

Virtual organization (VO) is a fundamental on sharing resources in grids. As the number of users and resource providers increases, also should the scalability of VOs to enable efficient utilization of resource sharing. This paper presents the scalable virtual organizations (SVO) for grids that federate VOs to share additional resources among VOs. VO agent is assigned in each VO to interact with the standard components of grid middleware and manages the procedure of requesting additional resources by another VO agent. To select appropriate VOs, we propose a group similarity function which is used to calculate similarities between VOs by using resource attribute tags as inputs and assist on deciding the appropriate VO. The throughput performance of using SVO is shown in our paper.