Research And Implementation Of P2P Search Engine Based On JXTA

Search engine solves the difficult problem of searching information. However, because traditional search engine adopts centralized mode, some problems still exist, such as server failure, limited storing capacity and outdated links that are not upgraded in time, which seriously affect performances of search engine.P2P has characteristics of distribution, dynamics and scalability. That P2P technology is to be applied into search engine brings new energy for search engine.How to bring P2P new idea and technology advantages to search engine is discussed in this thesis, and main content to research and problems to solve are the followings:(1) Because existing P2P applications are developed from the bottom with no standards, all of them can not be compatible to each other. Then, while designing, P2P platform—JXTA is chose as the development standards for P2P network, and a basic P2P communication network is built based on JXTA protocols.(2) Resource discovery is a difficult point in P2P network. IP multicast is used for broadcast searching within firewall, and HTTP is realized to search through firewall. Meanwhile, search peergroup is defined with membership service, and communication flow is limited within peergroup, avoiding needless communication.(3) Because searching results come from several peers, second sorting module is used to gather and sort results before displaying them to users. Considering dynamic character of P2P and user demands, based on Lucene sorting mechanism, second sorting mechanism is advanced to adapt to search in P2P.(4) Peergroup management service, pipe communication service, message management service, content download service and local resource management service above P2P network are defined, user-friendly interface is designed, and at last, one integrated P2P search engine based on JXTA is built.At last, implementing schemes of P2P search engine based on JXTA is present, then it is tested and analyzed in LAN. The experimental results approve that this system can deeply mine information stored in the computers lying on the edge of the network, and can sufficiently take use of the computing and storing capability, so it has great value of utility and foreground to widely spread.