SDRD: A novel approach to resource discovery in grid environments

Abstract

Scalability and routing efficiency are two crucial aspects of resource discovery in grid environments. SDRD, a novel approach to resource discovery in grid environments, is proposed in this paper. Basic model of resource discovery is proposed. Based on the basic model, formal description of SDRD is given. Sophisticated routing strategies are implemented in SDRD to improve the scalabilities and routing efficiencies of existing peer-to-peer based discovery approaches to grid resource. Improved Distributed Hash Tables (DHTs) are adopted within virtual organizations (VOs) to achieve better local routing efficiency. SuperNodes are only in charge of forwarding requests to other Super-Nodes in different VOs when the desired resources are not present in local VO, which enhances the scalability due to the lower load on SuperNodes. SDRD could be applied to other resource management models to offer excellent scalability and high routing efficiency. Results of analysis and simulations show that SDRD scales well and could offer an efficient decentralized discovery service in grid. © Springer-Verlag Berlin Heidelberg 2007.