Peer-to-peer full-text keyword search of the web

Abstract

Full-text keywords search of the Web over structured peer-to-peer networks shows promise to become an alternative to the state-of-the-art search engines since P2P overlays propose means for decentralized search across widely-distributed document collections. However, a disadvantage of structured P2P systems is that they consider only the problem of searching for keys, and thus cannot perform content-based retrieval. To deal with this problem, in this paper we consider a full-text retrieval problem in structured P2P networks. Our keyword searching engine BI-Chord is build on top of DHT-based P2P systems, entirely distributed, uses bloom filters and inverted index and therefore scales well with the size of the network. Experimental results show that our mechanism is efficient, scalable and provides high quality of search results, i.e. the precision and recall metrics.