Font Size: a A A

Research On P2P-based Crucial Technologies For Web Services

Posted on:2011-08-24Degree:MasterType:Thesis
Country:ChinaCandidate:J LiFull Text:PDF
GTID:2178360308465547Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Web services is network-based and distributed modular component, which defines how applications can achieve interoperability in the Web, and it has become a distributed network integration application technology with potential. Web services discovery is the crucial technology among the Web services applications, and its goal is to find an appropriate Web services, which is to locate a disposal services description in machines that is previously unknown and meets certain functional standards. Organizing publishing and discovery architecture for Web services, and then quickly finding the expected Web services are an important prerequisite for reusing and compositing the Web services. The popularity of the application makes the number of Web services increasing rapidly and updating information frequently, and it will lead the traditional centralized UDDI-based Web services discovery system to face many problems: a single point failure of the entire UDDI server system because of the huge amount of data, highlighted performance bottlenecks, difficult maintenance, low flexibility and so on. Resolution of these issues will have great significance to the current Web services technology.P2P (Peer-to-Peer) is now a kind of popular distributed network model. This model has the characteristics of parallelism, dispersibility and independence and breaks the tradition of C/S model, and each node in the network is equal, so it can effectively solve the bottleneck and single point failure of traditional Web services system if combines with Web services. Therefore, building publishing and discovery systems of Web services based on P2P technology can effectively make up for the structure shortages of the traditional centralized UDDI architecture.This thesis presents a two-layer Web services publishing and discovery system combining distributed and centralized features, in which registry center looks like a centralized organization, but in fact allows multiple registries to coexist in the network. This method decentralizes Web services registry center using Web services technology and P2P technology. The bottom layer uses P2P networks to connect services nodes, and each network has a super node as the local UDDI registry center, which organizes and registers services information independently; The top layer Web services provider publishes and registers information to global UDDI registration center, and it is transparent to participants during the services publishing and discovery process in the entire network.Based on the two-layer Web services publishing and discovery system, we also design organizational structure and registration process of registered Web services nodes in Web services publishing and discovery system in distributed environment.For lack of quality of services (QoS) description in typical Web services discovery architecture, it is difficult to select the best services from a mass of Web services being of the same function for the users. So QoS is put forword and quantified for services, which enables the Web services requested to meet the users'non-functional requirement, in the same time allows users to feedback their QoS values of the services that they use, so it can improve the Web services credibility. And on the basis of it, we achieve QoS-based Web services matching process in the two-layer architecture.In addition, we put forward a cluster analysis method thus K-Wmeans algorithm which adapts weighted improved strategy aiming at dynamic and distributed features of P2P networks. In this method, all Web services on the nodes have been distributedly clustered, so it enables services requesters to find the required Web services by category more quickly and accurately and then improves search efficiency. We analyze the time complexity of the algorithm, and prove that it has superiority of performance by experiments while applying this method to the current Web services clustering.We can enhance flexibility and scalability for Web services discovery architecture using discrete technology of P2P, solve performance bottlenecks for single point of failure in network effectively and make the network be better robustness, and then improve efficiency of services publishing and discovery. These are very important for the development of Web services technology as well as P2P technology.
Keywords/Search Tags:Web services, P2P, Web services registration and discovery, QoS, K-Wmeans
PDF Full Text Request
Related items