Research On The Distributed Query Algorithm Based On P2P System

Peer-to-peer (P2P) is a network architecture which resource (computing, storage, communications and information, etc.) are distributed use and sharing in it, and it is corresponding to the Client/Server (C/S) architecture that currently occupied a dominant position in the network. Basing on the Web applications, the C/S structure has a great success. However, in this architecture, the network capacity and resources are concentrated in the central Servers, which become bottlenecks in open network and capacity expansion. Contrary to C/S network architecture, P2P network infrastructure has no central node on media communication, and the nodes (Peer) are equally, that is, each node can be carried out equity communications and has media content receiving, storing, sending, integrating and searching the media metadata and the searched functions. In P2P network architecture, the advantages are the P2P network node capability and resources can be shared, in theory, network capacity and resources is the sum of P2P nodes. In the P2P architecture, contents are no longer concentrating on the central Servers in the network, instead of being distributed in the P2P nodes on the edge of the network around users. P2P technology makes business applications evolution from a centralized system to the distribution, in particular the distribution of servers, which overcome the bottlenecks in business nodes and greatly reduce the construction and use costs of system, improve network and equipment utilization.In a typical P2P network, the data resource distribution in various independent nodes, how to indexing, searching, locating and accessing these data and information resources efficiently is an important issue. The latest results are distributed search and routing algorithm based on DHT (Distributed Hash Table). In the application layer DHT organized all the P2P nodes into a structured overlay network that file indexes in it, and message enquiries will be routing through the overlay network. Through distributed hash function DHT mapped the entered keyword to an only node in the overlap network, and then to connect with the node by certain routing algorithm. In P2P network typical DHT topology models are CAN, Chord, Pastry, Tapestry and so on. In Chapter 2 the various DHT model are described in detail.At present DHT faces many problems, the biggest problem is in the initial design DHT overlooked the nodes vicinity in the physics network, resulting in the link between overlap network and physical network is small, that is DHT failed to take full advantage of the underlying physical network topology information. This locating resources method will affected the localization of the data, through DHT a local data maybe mapped to a node of its long physical distance from the local node, thereby causing actual finding inefficiency, and the cost of resources locating greater. Because the routing algorithm is the core of DHT, improving finding efficiency based on DHT is currently the focus of the P2P study, it will be very important significance. Based on the above information technology, combining the new characteristics in P2P network with the specialities of positioning resources, around DHT finding efficiency improvements, the optimizition scheme are put forward and the the performance analysis is carried out. By analyzing these programmes are clarified to effectively improve the efficiency of existing DHT routing.The main work and contributions are as follows:1. Bidirectional routing structure in Chord ring is proposed. In the routing table in Chord based on the DHT, there is unidirectional clockwise routing along ring. Key locating is carried out only clockwisely, when the destination node fell into the latter half ring from the beginning of the current node clockwisely, the hops along anti-clockwise maybe fewer than along clockwise. An improved method is presented which is bidirectional routing table. Routing can perform along clockwise and anticlockwise according to the locations of the current node and destination node. The next hop direction is optimization that the next node is the nearest one apart from the destination node. This strategy reduces the average hops and enhances the search efficiency.2. Put forward the DHT routing table structures base on the B+ tree. Because B+ tree is sorted balanced multi-branches tree, the routing structure in B+ tree really sets up tree-index routing on nodes in order to achieve the scope search. The node identifiers with B+ routing structure can narrow the search scope, reduce the hops and accelerate lookup speed, and it can control the lookup length in the height of B+ tree. 3. IPv6-based distributed hash table is brought forward. As IPv6 address are hiberarchy and aggregation, routing aggregation will be done by IPv6 address hiberarchy allocation to resolve the issue of the overlap networks divorcing from a physical network. Routing latency in the network is mainly between domains, and the delay in the region is relatively small, so if want to reduce enquiries time, it is necessary to reduce the inter-domain hop. In the network nodes using IPv6 address, the node identifier with the hierarchical structure is built through hashing the different parts of the IPv6 address, so that the nodes in the same domain or neighbors nodes can also near each other in the overlap network, and logical network is more matched with physical network, and logical hops inter-domain to be reduced, greater efficiency to be improved.4. A P2P database model based on Chord is set up. After above three points of improvement in Chord a P2P system database model based on Chord is constructed - LRM (local relational database model). It is in order to solve the problems that bottleneck node and higher network transfer load for keeping the data consistency between nodes are in distributed database system based on client/server. An instance based on LRM analyzes data transfer by domain relation and indicates the way which realizes coordination formulas between P2P nodes. By instance, lower communication, more flexible sharing on data and service and better reliability are proved to be the attributes of the LRM.