【作者】 周玉姣；

【导师】 钟诚； 李智；

【作者基本信息】 广西大学 ， 计算机软件与理论， 2008， 硕士

【摘要】 Web服务将程序封装成单个实体并发布到网络上以供其他程序使用的功能,可以使企业和个人迅速且廉价地向网络用户提供服务。Web服务具有十分广阔的市场前景。随着Web应用的迅速发展,Web服务的数量急剧增长,如何从众多的Web服务中找到真正满足要求的服务已经成为目前Web服务研究领域重要的课题。当前主流的Web服务发现方法是基于UDDI框架的服务匹配的,由于UDDI注册中心仅提供基于关键字的查询,缺乏对语义描述的支持,服务匹配仅仅是基于关键字的匹配,不能实现基于服务功能的语义匹配,从而造成服务发现的精确度较低。另外,当前的UDDI注册模型是一种集中式架构,当注册服务的数量急剧增多时,面临系统瓶颈、单点故障、查找效率低等诸多问题的影响和制约,从而严重影响了Web服务发现的效率。本文在研究P2P环境和基于语义Web服务发现基础上,以获得高效率的Web服务发现方法为目标,考虑实际网络中节点能力的差异,引入超级节点,结合One-hop路由和Chord路由机制,构造了一种双层环P2P结构路由,并设计相应的区间查询定位和消息广播算法,提出一种基于本体聚类的双层环P2P网络Web服务发现方法。该方法利用基于双层环P2P网络组织Web服务的注册节点,用服务领域本体概念映射区间对注册节点按其提供服务的类型进行聚类,使特定领域的Web服务的注册与发现集中于双层环P2P结构的外环少量的连续节点上;充分利用双层环P2P网络中搜索和定位节点的动态机制,使服务查询请求能迅速从代理节点转发到相应类型注册节点中。分类中的每个注册节点之间是对等和自适应的且没有严格的相互依赖关系,注册节点的负载得以减轻,并能有效提高查询效率。实验结果表明,本文提出的的方法可以应用于注册节点数目较大的情形,并且在网络规模增大时仍能保持较好的性能。更多还原

【Abstract】 Web services will provide the packaged applications as a single entity and publish to the network to other procedures for the use of the capabilities. So enterprises and individuals quickly and cheaply can publish their digital assets to network users. Web services are very broad market prospects. With the rapid development of web applications and the rapid growth of web services’ number, web service discovery is one of the burning questions in the area of web service research .It is the important component of web service system.Web service discovery is one of the burning questions in the area of web service research.It is the important component of web service system. However, the present main-steam web service discovery method is based on the service matchmaking within the UDDI framework, the UDDI register center can only offer the keyword-based inquiry, and cannot achieve the semantic matchmaking based on service function, a low rate of inquiry precision is resulted in. Furthermore, the presented UDDI register model is a centralized architecture, when the number of register services explodes, it will cause such problems as bottlenecks, single points of failure, and low inquiry efficiency, then the efficiency of web service discovery will be greatly affected.In order to achieve the classification of web services register nodes and exact location of query request, TPROC (two pass routing on chord ),a new lookup algorithm is constructed by combining the lookup protocols of One-hop with Chord and taking into account the ability differences of nodes in the network systems. The interval query locating and message broadcasting algorithms are designed, a web service discovery method based on ontology clustering in TPROC networks is presented. One of its features is classifying service registry nodes according to service ontology on outer ring of TPROC networks. So that specific areas of Web services request can be processed in small service registry nodes. Other feature is that the Acting node can quickly transmitte query request to the appropriate type of service Registry node by using full TPROC networks in the search and the dynamic positioning system. By doing these, the network overload can be alleviated and the efficiency can be improved. The experimental results show that the proposed approach can be applied to the case for larger number of service registry nodes and its maintain good performance can be maintained with increase of the size of the P2P network.更多还原