Research On Unstructured P2P Network Searching Mechanism Based On Interest Similarity

P2P network is also called peer to peer network,which is mainly used in the fields of file sharing,instant messaging and resource search.Compared with the traditional client / server mode,the P2 P network can reduce the dependence on the central server,and has the advantages of centralization,scalability,load balancing and so on.All nodes in the P2 P network are peers,which can be used as a network service requester and respond to requests from other nodes in the network.Unstructured P2 P network has been paid more and more attention due to its simple topology,easy implementation of resource search and less network constraints.Unstructured P2 P network resource search algorithm is mainly based on improved flooding algorithm,but these algorithms have some obvious problems: one is the central idea of the blind search algorithm,the search efficiency is not high;the two is the search process will produce a message sent to a large number of redundant messages from neighbor nodes,easy to cause network congestion.In this paper,for overcome the shortcomings of the resource search algorithm in unstructured P2 P networks,the following three improvements are proposed.1.In this paper,a new structure of a binary search tree topology based on interest similarity for unstructured P2 P networks are proposed.The construction of a binary search tree node is based on similarity of interest between the content and the nodes whose the value of similarity of interest is high will cluster into a binary search trees.In a binary search,the similarity of interest of the node of the left is less than the threshold value of the root,and the similarity of interest of the node of the right is more than the threshold value of the root.Due to the characteristics of the a binary search tree,the algorithm can get better search success rate in the process of resource search,and will greatly reduce the redundant messages in the search process.2.Since the efficiency traditional colony search algorithm is low and relies too much on the path information concentration which will lead to the premature convergence problem,in the process of search this paper will consider credit evaluation,and will introduce the pseudo random proportional rule,so that the resource search process can access to more nodes.By this way premature falling into a local optimal solution will be effectively avoided.3.Based on improved ant colony algorithm in unstructured P2 P network in the fourth chapter,a search algorithm for unstructured P2 P network based on the binary tree ant colony algorithm is proposed.The node in binary tree make own resources stored in the parent node,so during the resource search process,only some nodes will participate in resource search and query forwarding,which reduces the forwarding nodes to a certain extent,and reduce the redundant messages in the network.