|
In recent years, with rapid development of computer technology, network technology and database technology, P2P technology once again receives close attention and is widely applied. Since P2P technology has such feathers as non-center, scalability, robustness, high performance/price ratio, privacy protection and load balance, it fits in file share, peer computation, cooperation and search fields. At present, P2P research work majorly focuses on network models, search problems and resource management.We firstly propose the semantic model of taxonomical data, and then propose Taxonomical Data Ring Network (TDRN), data in which follow a semantic model. In TDRN model, we provide Simple Routing Mechanism and Complex Routing Mechanism. In TDRN model, we provide peer join and leave operations and item search, insert, update and delete operations, and these operations satisfy common needs for file share in P2P networks.In our implementation we substitute the simulating TDRN model for real one in environment of a single computer. We majorly design and implement algorithms of peer join and leave, and item search, insert, update and delete. Algorithms of item insert, update and delete are based on item search, which includes domain search and exact search. Later we carry out theoretical analysis on time complexities of these algorithms under the two routing mechanisms.In experiments, we generate a three-layered tree and a four-layered tree as semantic models of taxonomical data, and then we generate items randomly, which are used for item operations. We measure the network size with the peer amount and the item amount. With the network size increasing, we test response time, message amounts and hops of the above algorithms. Although changing of response time in simulating TDRN is not the same as that in the real one, message amounts and hops in simulating TDRN are in consistency with those in the real one. Results show that TDRN model can effectively deal with taxonomical data.
|