【作者】 寇月；

【导师】 于戈；

【作者基本信息】 东北大学 ， 计算机应用技术， 2005， 硕士

【摘要】 随着Web服务的发展,“网格”概念的提出为基于Web服务的企业柔性业务集成开辟了新的途径,目的是试图实现互联网上所有资源的全面贯通,彻底消除资源孤岛。在此基础之上,又提出了“服务网格”的概念,它呈现给用户一个整合的虚拟服务资源仓,并实现资源智能化自我管理,使Web服务的功能和形式更加丰富,满足更多元化的个性服务需求。 由于服务网格应用的普及,网格服务的数量急剧增加,面对如此数量庞大的服务群,如何对其进行组织,怎样能够从中找到用户真正需要的服务,就成了一个十分关键的问题。需要有一种机制来对服务的信息进行统一管理、提供对网格环境的一个实时的动态反映,这就是元信息管理机制。为了能够快速、准确地查找性能最优的网格服务并对其状态进行实时监控,文中对现有UDDI系统的查询方式、存储形式和安全等方面进行了改进,建立了一个基于层次的支持网格服务的元信息管理机制,来管理网格服务的发布、发现以及信息监控与更新。 本文,首先对已有元信息管理机制进行了深入分析和探讨,提出了一种层次的元信息管理机制,有效地实现了服务资源管理。该元信息管理机制依据网格服务实现的功能对网格服务进行分类管理,即将网格服务分为资源级服务和集合级服务两类。同时基于分布式的思想,根据网格服务的分类将元信息管理机制划分为三层。该基于层次的元信息管理机制,可显著地提高服务查询的查全率和查准率,并且具有分布性、可靠性和可扩充性等特点。 在此基础上,本文进一步探讨了利用语义信息和服务描述信息智能化地选择功能最匹配、性能最优服务的解决方案。具体分三步实现:首先对服务的注册信息和查询请求信息进行标准化处理;然后根据语义信息确定服务所在的组织和域,并确定满足功能需求的服务;最后依据服务约束信息,基于服务约束匹配算法选择最佳的服务集。 通过实际验证,本文所提出的网格服务内基于UDDI的元信息管理机制和资源发现策略是可行的,达到了预期的目标。更多还原

【Abstract】 With the development of the Web service, the presentation of grid breaks a new path for enterprises’ flexible compositions. The purpose of grid is to implement the overall sharing of resources on Internet and eliminate the island of information completely. With the advance of grid, service grid is the new concept combining Web service and grid. And it provides a virtual register of resources and realizes self-control intelligently, which makes the function and format of Web service more plentiful and satisfies personal requirements much more.Because of the popularity of service grid and the increase of grid services, it is crucial to resolve the problems including how to organize so many grid services and how to make users discover appropriate services. A mechanism is needed to manage services uniformly and supply a dynamic reflection of grid environment. Such is meta-information management mechanism. In this paper, present UDDI in the way of querying, storing, security, and so on are improved in order to enhance the ability of discovering appropriate grid services with high performance efficiently and real-time monitoring the dynamic information of services. A multi-level meta-information management mechanism to manage grid services on registering, discovering and monitoring is proposed.First, we have carried out deep analysis and research into current meta-information management mechanisms and have presented a new layered one which realizes management for service resources efficiently in this paper. The mechanism enforces classified management for grid services according to their different functions. And they can be classified into two classes: resource service and collective service. Based on the distributed idea, we classified the meta-information management mechanism into three levels. The layered architecture can enhance the efficiency of inquiring services greatly and has some characters including distribution, reliability and extensibility etc.According to the semantic information and services’ descriptive information, an intelligent solution of discovering the most matching services in function and selecting the best service from them in performance is also addressed. This process can bedivided into three parts: Firstly, both the information of services and the requests of users should be converted to a standard form of system. Secondly, the appropriate organization and domain where the service is located can be determined and we can find the services that satisfy the users’ requirements in functions. Finally we can select the most matching services according to a matching algorithm of restriction.By the use of this meta-information management mechanism presented in this paper, we have validated the feasibilities of some information management policies.更多还原