Study On Web Service Discovery Mechanism In P2P Network

With the boom of P2P network,how to find out the needed services becomes the most concerned problem in web service technology field. However, there are a lot of problems in traditional service discovery mechanisms. All the popular discovery algorithms as flooding, random walk and Gnutella2 have both advantages and disadvantages. How to take care of discovery speed and accuracy rate and keep steady system cost at the same time, this is the root question of P2P searching algorithm.On the base of P2P network model,"Small World Phenomena"and rumor spread mechanism, this paper puts forward the P2P e-commerce service discovery model MatureNode.At first, this paper imports the conception of"Mature Node". Using"Node Mature Metric (NMM)"to make every peer owns some kinder"instinct"and calculating NMM according to three different"service information bases". Using OPNET simulation to compare with random walk algorithm, this algorithm has better efficiency.The next step is simulating the process of rumor spreading to improve MatureNode model and built up a new ad forwarding mechanism. This ad forwarding mechanism contains two key parameters:"Forwarding Probability (Pf)"and"Forwarding Weight (Wf)". Pf determines the number of neighbor nodes ads forwarded. It has inverts radio with ad's popularity, so rare ads will spread quickly and widely. Wf decides which group of neighbor nodes is chosen to forward ads. According to the results of simulation, on the base of steady system cost, MatureNode can obviously improve service discovery speed, furthermore it can discovery more services in provided TTL.