A Self-Organized Grouping (SOG) Framework for Efficient Grid Resource Discovery

Dynamic and heterogeneous characteristics of large-scale Grids make the fundamental problem of resource discovery a great challenge. This paper presents a self-organized grouping (SOG) framework that achieves efficient Grid resource discovery by forming and maintaining autonomous resource groups. Each group dynamically aggregates a set of resources together with respect to similarity metrics of resource characteristics. The SOG framework takes advantage of the strengths of both centralized and decentralized approaches that were previously developed for Grid/P2P resource discovery. The design of SOG minimizes the overhead incurred by the process of group formation and maximizes the performance of resource discovery. The way SOG approach handles resource discovery queries is metaphorically similar to searching for a word in an English dictionary, by identifying its alphabetical group at the first place, and then performing a lexical search within the group. Because multi-attribute range queries represent an important aspect of resource discovery, we devise a generalized approach using a space-filling curve in conjunction with the SOG framework. We exploit the Hilbert space-filling curve’s locality preserving and dimension reducing mapping. This mapping provides a 1-dimensional grouping attribute to be used by the SOG framework. Experiments show that the SOG framework achieves superior look-up performance that is more scalable, stable and efficient than other existing approaches. Furthermore, our experimental results indicate that the SOG framework has little dependence on factors such as resource density, query type, and Grid size.