An Experimental Study of the Active Recommendation Mechanism Based Distributed Approximate Indexing in Unstructured P2P Networks

In unstructured P2P systems, it is hard to find the desired files without distributing queries, since there is neither a centralized directory nor any precise control over the network topology or file placement. Many valuable efforts have been done towards improving search efficiency and increasing the scalability of unstructured P2P systems while guaranteeing endurable resource consumptions. In this paper, we purpose an active recommendation mechanism (ARM) which aims to provide efficient searching services and enhance benign competitions in P2P networks. We also build a two-tier proxy architecture consisting of reputation evaluating proxies and content indexing proxies. Based on the ARM and the new architecture, we purpose a distributed approximate indexing (DAI) method to support an efficient search without the need for a single centralized index host