Research And Application Of P2P Search Model Based On Chord

The emergence of search engine has brought great convenience to users in retrieving on the Internet. But the traditional search engine also showes deficiencies as the further development of the network and more quickly the content of Internet updates. And the distributed network based on P2P, possesses scalability, robustness, load balancing, etc., compared with traditional distributed systems. P2P technology has the unparalleled advantages, and is suitable for building distributed information retrieval systems to implemnt the local information on the computers.We firstly proposed an improved routing algorithm RChord based on Chord by adding a neighbor node list and a node cache table to Chord's routing table. The neighbor node list provides nodes adjacent to the node in physic, and the node cache table provides the hot nodes. Both of them can improve the efficiency of searching.Then combined with the improved RChord algorithm and unstructured P2P network Gnutella, P2P technologies and search engine, we proposed a two-tier distributed search engine model based on P2P, and optimized the behavior of super-nodes in this model. Each host is a node, and super node with strong performance, and the others are the common nodes. Super-nodes form the Chord model and are responsible for positioning the index nodes containing keywords related to super-node information, and then forwarding queries to common nodes. The control of super-nodes in the model is very important, and we proposed a control strategy of the cache which records the candidate super-node to substitute for the leaving super-node or super-node with low performance.Finally, we implement a confidential inspection system based on the application model. Use the model proposed, we design and implement a campus confidential inspection system, to check the confidential information on the end hosts.