Search Mechanism On Unstructured P2P Systems Research

Along with the network technology development, changing the life of people, makes us more and more cannot leave Internet, along with the popularization of broadband is not confined to the traditional Internet chat, download, read these, but also hope that through the network of real-time voice, video watch. Unlike the C/S mode, P2P it is a peer to peer network structure:the nodes in the network can be used as a client or server nodes exist, between each other and their respective data needed, and unlike the C/S mode, the data must contact with SERVER if you want to get. Through this model, not only greatly reduce the pressure of server, save the bandwidth, but also make the full use of the bandwidth. This kind of network structure in stream media application is very obvious.P2P network is a distributed network form, it does not distinguish between server and client, which is functionally consistency. A P2P system file is not stored in the central server, but stored in the dispersion of the terminal nodes, file transfer is directly among the terminal nodes. The system does not have a centralized index directory, so how to effectively to locate to resource becomes the core problem in the study of P2P.The current P2P system has3kinds of model:a fully distributed (centralized directory), pure distributed unstructured, based on distributed hash table(DHT) distributed structured model structure. This thesis mainly focuses on discussing the unstructured P2P cyber source search, as long as a pure distributed unstructured network resource search problem, in the fourth chapter, a detail study of the ant colony algorith (AOC), Providing two kinds of improved ant colony algorithm, by setting the suspicious coefficient δ, used to judge the ant path selected would constitute a region optimal solution. In the ant colony algorithm late period,using flooding search mechanism to improve search efficiency, accelerate the convergence rate, reduce the search time and relay information. In the fifth chapter based on the similarity of structure of resource location and search algorithm.Different resource types of similarity are classified not by the node similarity to classification, calculation each node in the similarity of various resources, form a list to record in the form of resource similarity information, also record the similar nodes’IP address to help positioning and searching the resource. this algorithm saves time, reduces the amount of flooding message.