SARIDS: A Self-Adaptive Resource Index and Discovery System

摘要

Recently, the resource sharing systems apply the P2P technique to provide scalable multi-attribute range queries. However, due to the heterogeneity of resources and the variation of sharing policies in different providers, current P2P-based resource discovery systems may suffer the load imbalance problem in a large scale distributed system. In this paper, we propose a self-adaptive resource index and discovery system (SARIDS) to achieve load balancing. SARIDS adopts a two-tier architecture based on the structured P2P overlay. The intra-overlay is constructed by normal peers with the same attribute via the locality preserving hash function; and, the inter-overlay is constructed by super-peers with classified attributes in different intra-overlays. SARIDS supports not only the multi-attribute range queries but also the self-adaptive mechanisms for load balancing in the intra-overlay and among the intra-overlays. The simulation results show that SARIDS is scalable and efficient for load balancing even in the nonuniform peer range environment.