Research On Improved Structured P2P Network Based On Chord Algorithm

With the development of Computer Network Technology, how to get useful information from a heavy volume of messages has become the focus of today’s network technology research. But in the traditional C/S mode, all resources are only stored on the server, the search must be worked on the server, the model leads to the central server has been under great pressure, which have been unable to meet the needs of users today. Therefore, P2P network technology let the problem of resource locating which is now faced with be effectively addressed. In P2P networks, as the term of P2P, the exchange of information between network nodes are all working together, relatively under equal pressure. However, the node’s physical information is not considered by the current structure of P2P networks. At the same time, because of the performance of the nodes is different, which make the searching be low efficiency, leading to a waste of resource.We conduct the following work:1. In this paper, in which the Chord algorithm model is focused on, In order to address the issue of mismatch between the physical topology and logical topology, which leads to the low efficiency, matching problem as a traveling salesman problem (TSP), uses the genetic algorithm to find the best solution of the problem, then constructs the Chord ring, which solves the problem of mismatch between physical topology and logical topology.2. The algorithm combining genetic algorithm regards the topology Based on the above, using the Bi-directional routing look-up mechanism to improve the efficiency. The experimental results show that the proposed algorithm is more effective than the conventional Chord-based method on the length of query path and the hop of query.3. Considering the offect of grouping mechanism in large scale network, aiming at the problem of low efficiency because of large numbers of nodes, presenting the DS-Chord algorithm based on physical domain. The algorithm let all nodes be grouped based on the physical domain, the nodes of each group in the physical distance is the shortest, so that to make the Chord ring logical topology and physical topology achieve the maximum matching. And then select a optional performance in each group as a supernode, supernodes of all the groups build a super Chord ring. When a node receives the query request, resource is not local, firstly searching the super Chord ring, finally going back to the common query. The DS-Chord algorithm gives full play to the performance of super nodes, at the same time, due to the shortest distance of physical node, therefore, when the number of nodes is huge, the problem which searching is low efficiency has been effectively solved.