Research And Implementation On P2P-Service-based Search Engine

Peer-to-peer technology has been wildly used over the Internet, especially for file sharing. People can hardly get what they need without search engine. P2P based search engines are different from web ones:costless, higher recall rate, etc.Analyzed some typical P2P based search method and found they were composed by specific architecture and three basic methods: flooding, indexing and hashing. These methods have both advantages and disadvantages. They are not sufficient for various using.Bringed up a P2P service based multi-level model to separate the P2P applications and P2P network topology. In this model, all P2P applications are composed by P2P service. The P2P service level can afford distributed deployment for P2P application level. This abilily drived from the Petri net. P2P services have two basic types: computing service and storage service.Designed a centrated index based search engine on P2P network. Implemented the interaction, deployment and load balance of P2P services. This article also bringed up a new multi-keyword search method: decision tree method for the greatly distributed P2P network.The P2P service based centrated search engine integrated the advantages of both P2P and centraed search engine. It provides a new method for search engine.