P2p Distributed Database Research

P2P network has robustness, adaptive, availability and load-balance, is one of next generation Internet technologies. With the development of Internet, the distributed database systems based on LAN have poor scalability, rigid architecture and require of maintaining data ACID, which can not afford to manage the data distributed among world-wide network. The P2P technology can make up the distributed database system's lacks and supply a effective network platform for managing and sharing data in Internet environment. The thesis designed an implementation mechism of P2P databse based on DHT (Distributed Hash Table) P2P network.DHT-based P2P network has robustness, adaptive, availability and load-balance but it lacks methods of description and management of complex semantics data, and supplies only exact-mach query bases on keywords.At first, the thesis analysed the distributed database system technologies, and compare it with P2P. Next, we analysed the P2P three kinds of architectures, and researched on the classic P2P data management system from network toplogy, architecture, data replica, etc. Based on all above, we proposed an implementation mechism of P2P databse based on DHT-based P2P network.The replica placement, resoures index and location mechanism of complex semantics data is the core of P2P Database. In this paper, a replica placement strategy and a mechanism of data Index and resources location in P2P Database was presented, which maintained complex semantics data representation. Data replicas placed on other P2P nodes improved the fault-tolerance and availability. A distributed hierarchical index mechanism was proposed, by which the system could locate resources effectively and supported complex queries. The mechanism of resources location based on DHT could bridge a gap between DHT P2P and complex semantics data management.A group of peers which have tigher relationship and exchange data frequently would be composed into a intrest group. Coordination rules were proposed, by which peers could exchange and share data dynamicly.Based on hierarchical index and DHT-based P2P routing algorithms, a resources location method was designed to locate data (relation table) and resources requested by user.