Ouster Mechanism Study Based On Indexing Network Of P2P Spatial Vector Data

Geographic information system has attracted more attention since the appearance of it, in thespan of a few years the popularization of user is rapidly developed. Of course, the development ofcomputer technology and the promotion of global Internet scope are inseparated. Now with the P2Ptechnology as a new distributed network technology for spatial vector data, the load imbalancedecreased and "hot point of bottleneck" problem solved in a new way. The traditional client/servernetwork model can be subverted by P2P technology, so that information can be free duplextransmission; the network customer resources can effectively used. The new way can be used toimprove the space vector geographic data index performance and organization efficiency, thedevelopment of geographic information system is led to a new milepost.In this thesis, existing technology using P2P spatial data indexing network is studied andsummarized, in the process of evolution of the geographic data indexing network progress, thecluster mechanism of spatial indexing is based on spatial indexing network have been learnt. Incomprehensive consideration of nodes’ delay and bandwidth on index network, the two metrics canbe balanced by an adjustable weight, then the cluster member node optimal joinment algorithm isrealized; the weights fit condition is used on cluster mechanism with the maximum group size, thendynamic member node exit and dynamic join is regulated, and the cluster member node dynamicreplacement algorithm is realized. By using Peersim simulation tool, the effectiveness of ourproposed algorithm is verified in the simulation platform. Finally, on the JXTA platform, P2Pspatial data indexing network prototype system is realized, the management node algorithm usingcluster mechanism is improved, and the system of the proposed optimization algorithm is tested.The test results show that the client’s resource is effectively used by the improved clustermechanism. In changing network environment, the node selection algorithm can be adjustedadaptively and effectively, the query efficiency can be enhanced.