Research On Search Technologies Unstructured P2P Networks Based On Local Information And Request Of Peer

Unstructured peer-to-peer network is one of key technologies used to improve search efficiency of large-scale information. However, it is very difficult to obtain global network information in unstructured peer-to-peer networks as the increasing of network scale and information content. In this case, this thesis deeply studies search mechanisms of unstructured P2P network based on peers’local information and requirements. The main work and achievements are as follows:(1)Reducing the network overhead generated during searching is important in the study of unstructured P2P networks. Flooding algorithm and random walk algorithm are simple and easily implemented. However, they generate a large number of redundant messages in the search process and consume much network resources. To solve this problem, an effectively limited search mechanism (RFSA) is proposed. The search path and redundant search path are defined in the RFSA. Repeat receiving and forwarding of messages are eliminated by restricted receiving of messages. RFSA uses information of real-time search path carried in the search process to select neighbor nodes that do not appear in the search path. Then the neighbor nodes will be used to forward query messages to eliminate redundant search paths. The number of messages and network overhead generated by RFSA are analyzed theoretically. In simulation experiments, comparative analyses of the limited search mechanism and non-limited mechanism flooding and random walks algorithm are done in terms of the network overhead, query hit rate, search coverage rate, and the number of redundant messages, etc. The results show that this method reduces the generation of redundant messages, and cuts down the network overload.(2) In unstructured peer-to-peer networks, the blindness of search causes a large number of network overhead. Heuristic search mechanism is an effective method which depends on the previous searches to guide future ones. In the existing methods, searching for high-repetition resources is more effective. However, their performances need to be improved when they are used to search for non-repetition or low-repetition or rare resources. For this problem, considering the similarity between social networks and unstructured P2P networks, we propose a credibility search algorithm based on different query according to trust generation principles in sociology and psychology. Firstly, we give a credibility calculation method to calculate neighbor’s credibility according to familiarity and similarity among queries and nodes. Secondly, queries are divided into familiar queries and unfamiliar queries. For different queries, different ways would be adopted to get the credibility of a node to its neighbors, and then neighbor nodes with higher credibility will be selected to forward queries. Experimental results show that the proposed method can improve query hit rate and reduce search delay with low bandwidth consumption in the three different network topologies whether they are under static or dynamic network environments. Futhermore this method has good adaptability to P2P dynamic environment.(3) Flooding algorithm and random walks algorithm are two typical search algorithms in unstructured peer-to-peer networks. Although flooding algorithm covers the most nodes, it generates a large amount of redundant query messages during message propagation. On the contrary, random walks algorithm can effectively reduce redundant messages. But it takes longer search time and covers fewer nodes. To addressing the problem, a two-stage hybrid search mechanism based on credibility of node service ability (CNSA) is proposed. In CNSA, neighbor node service ability is defined based on the local information of each node and a new neighbor node selection criterion is given. At the same time, CNSA is divided into two stages. The first stage is a limited flooding with low TTL hops to spread search to a large scope. Heuristic random walks mothed based on node service ability is executed with high TTL hops in the second stage. CNSA algorithm makes full use of the advantages of flooding algorithm and random walks algorithm, and combines the advantages of heuristic search mechanism. CNSA can improve the broadness of search, reduce the blindness of the search, and improve the performance of the search.(4) In unstructured P2P networks, the existing search protocols are effective for popular resources, but searching for scarce resources is inefficient. Improving copy rates of scarce resources are the main method to solve the problem of search inefficiency. The query hit rates on scarce resources are lower when copies of scarec resources are small. So the existing passive copy replication strategies based on the success query are not suitable for improving popularity of rare resources. To solve this problem, an active replication and search strategy of scarce resources is proposed. In the search process, nodes with scarce resources actively initiate the search for scarce resources. And local demand information is effectively obtained in the process of search, and then scarce resources are copied to the peers that have demands for the scarce resources. The method implements the on-demand replication of scarce resources and improves popularity and query hit rates of scarce resources. Based on local requirement information, we provide three different kinds of on-demand replication strategies and a rare resource search algorithm. Experimental results show that the active replication and search strategy of scarce resources can effectively increase copy rates of scarce resources with lower replication consumption and network overhead, and then improve the query hit rates of scarce resources.