节点文献

Web服务发现若干关键技术研究

Research on Some Key Technologies of Web Services Discovery

【作者】 胡建强

【导师】 邹鹏;

【作者基本信息】 国防科学技术大学 , 计算机科学与技术, 2005, 博士

【摘要】 开放的网络化应用和面向服务计算(Service-oriented Computing)导致Internet环境下软件系统的主要形态、生产方式、运行方式和使用方式发生了巨大变化。软件系统正处在一个由软件服务实体组成的开放协同的软件环境中,未来网络软件的一种趋势表现为通过服务发现和服务动态组合构造软件应用系统。以Web服务为代表的软件服务及软件服务协同已成为一种新型的Web应用形态,如何实现服务按需发现,以保证服务有效复用、相容组合,对提高网络软件生产效率,实现跨组织的业务集成具有极其重要的意义。目前的Web服务发现在服务描述语言的全面性、匹配算法的灵活性和对服务组合支持的有效性等方面存在不足,是当前面向服务计算领域面临的一个重要挑战。为应对这一挑战,本文首先围绕一个完整的Web服务发现过程,深入研究Web服务描述语言、服务匹配算法、服务选择和系统拓扑等关键技术,然后建立一个Web服务发现系统的体系结构,从而为实现Web服务之间正确、灵活的交互以及服务动态组合提供更有效的支撑。本文的主要工作包括: (1) 在Web服务描述方面,针对现有Web服务描述缺乏全面性描述的不足,在系统研究Web服务能力描述语言的基础上,提出了通用Web服务能力描述模型{S,C,P},并且设计了一种基于服务质量的轻量级Web服务描述语言QWSDL(Qos-based Web Service Description Language),较全面描述了Web服务的功能、行为以及服务质量。 (2) 在服务匹配算法方面,将现有的服务匹配算法归纳为“三层次,五类型”的匹配模型。针对现有Web服务能力描述的匹配算法缺乏灵活性的不足,引进相似函数来度量松弛匹配的服务相似程度,在保证较高查准率和查全率的同时,较大地提高了服务匹配效率。针对Web服务流程执行描述语言因过于复杂而难以直接匹配的问题,引入标识确定性有限状态机ADFSA(Annotated Deterministic Finite State Automata)模型,将该问题简化为可到达路径有限状态机的匹配问题,从而降低了服务匹配的复杂性并有利于避免潜在的Web服务组合不相容。 (3) 在服务选择方面,将满足服务功能和流程行为需求的服务选择问题建模为多属性决策问题,提出了多维服务质量属性的决策模型DQos。针对主、客观赋权模式确定权重系数的局限性,提出了主客观结合的赋权模式,以更加自然的方式克服单纯使用主、客观赋权模式的偏面性,为选取优质Web服务以及筛选劣质Web服务提供重要依据。 (4) 在系统拓扑方面,针对集中式拓扑存在可扩展性差的不足,提出了一种基于双层结构的分布式系统拓扑Stratus。在处理查询请求时,提出了适合Stratus的二阶段查询传

【Abstract】 The open Internet application, together with the service-oriented computing, leads to great changes of Internet-based software system in main shape, production mode, execution way, and occupation mode under Internet environment. Application system is under an open coordination software environment composed of various software service entities, and one of the trends that Internet software will behave in the future is to construct software application system through service discovery and service dynamic composition. Software service and software service coordination represented by Web service have become a new Web application shape, and how to make service discovery effectively guarantee service reuses and composition compatibility are of great importance both in the improvement of Internet software production efficiency and in the implementation of cross-organization business integration.However, the current Web services discovery has problems in all-sidedness of service description, flexibility of matching algorithm, and validity of supporting service composition, which are challenges in the field of service-oriented computing. Facing these challenges, this dissertation firstly deeply studies some key techniques including Web service description language, service matching algorithm, service selection, and system topology, around an intact Web service discovery process. Then, it establishes an architecture of Web service discovery system to provide more effective support for the realization of correct and flexible interaction, as well as service dynamic composition among Web services. Main works of the dissertation are as follows:(1) For Web service description, the dissertation not only puts forward a common Web service capability description model{S, C, P}, but also designes QWSDL(Qos-based Web Service Description Language), on basis of systematically study on Web service capability description language, which gives a full description of Web service functionality, service behavior as well as Qos (quality of serivce) information.(2) For service matching algorithm, the dissertation presents a "three levels, five kinds" matching model by inducing the existing service matching algorithm. Since the current Web service capability matching algorithm is not flexible enough, the dissertation introduces similar function to evaluate service similarity of relaxed match, which on one hand, ensures high recall and high precision of service discovery, on the other hand, improves the efficiency of service

  • 【分类号】TP393.09
  • 【被引频次】39
  • 【下载频次】1303
  • 攻读期成果
节点文献中: 

本文链接的文献网络图示:

本文的引文网络