Research On P2P Search Algorithm Based On Rumor Propagating Mechanism

Allowing file exchange among the end client peers through Internet is the initial purpose of P2P (Peer-to-Peer). In the short history, P2P application has become one of the main application types that consume a large fraction of Internet traffic. The capability that P2P architecture supports massive users makes it suitable for rapidly deploying powerful and large-scale distributed applications with low cost.In order to take full advantage of a variety of resources in P2P networks, the most important thing is to find those resources effectively. However, because P2P network is dynamic and scalable, it is difficult to design an excellent search algorithm for P2P. Rumor or Gossip algorithms are a class of algorithms for message routing. They are also called epidemiological algorithms, since messages are spread in a network much like a disease in a susceptible population, that is, the neighbors to which messages are forwarded by each node are chosen at random. Comparing with the flooding algorithms, the rumor algorithms can decrease the cost in large quantity. But because of the randomness of the message spread, the algorithms can not assure the coverage ratio of spread, and increase the time that message propagates.In order to overcome the shortage of rumor algorithms, the author improved the traditional Rumor algorithms by utilizing the Gnutella's Power-law characteristics. Instead of selecting neighbor at random, the improved algorithms propagate message based on the degree of node, which overcomes the limit of selecting neighbor at random, and thus optimizes the performance of Rumor algorithms. The feasibility and affectivity of the improved algorithm are proved by using Gnutellasim, a type of the simulation software for Gnutella networks, to simulate the improved and the original algorithms.The main research work is as follows:(1) Introducing some basic notions and knowledge of P2P networks, including the definition and development stages of P2P, the advantages and disadvantages between P2P and C/S networks, and the applications and current development status of P2P.(2) Inducing the evaluation criteria for P2P search technology, analyzing the flooding algorithm that is used in P2P, and introducing several main P2P search methods, such as iteration flooding, DBFS (Directed Breadth First Search) and local index.(3) Improving the existing search algorithm based on the rumor propagating mechanism by combining current search algorithms, depicting the improved algorithm in detail, including the concrete implementation steps, the proof of the correctness, and the validation of the feasibility and effectiveness of the improved algorithm by simulation.