Facilitating Effective Resource Publishing and Searching in DHT Networks

Abstract

As a structured overlay network, DHT (Distributed Hash Table) has been widely used in P2P systems. Existing resource publishing algorithms, including One-step publishing in eMule, do not completely exploit features of DHTs and incur heavy network load. In this paper, we propose a novel resource publishing algorithm: Step-by-step publishing, in DHT networks. It is more effective and scalable with lower network load than the existing publishing algorithms. We also utilise the techniques of vector space model and user relevance feedback to optimise the Step-bystep publishing algorithm, which effectively improves search hit ratio and reduces network load. After analysing the dynamics of P2P systems, we present index maintenance schemes under highly dynamic P2P networks. Based on real datasets from Maze system, simulation experiment results show that this proposed approach has low network overhead and publishing cost, high search and download hit ratio.