【作者】 李立；

【导师】 杨放春；

【作者基本信息】 北京邮电大学 ， 计算机应用技术， 2007， 博士

【摘要】 下一代网络是一种可以提供话音、数据和多媒体等各种业务的综合性的开放网络，代表了电信网络的发展方向。随着网络的发展和不断涌现的新的业务需求，下一代网络在网络能力开放之后再进一步开放业务能力，并在此基础上实现下一代网络的资源共享成为网络技术发展的趋势。然而目前对于下一代网络资源共享问题的研究还很匮乏。以OGSA为代表的网格技术的出现为研究解决下一代网络资源共享问题提供了较好的途径。因此研究如何借助于网格技术解决下一代网络资源共享问题将是一个有意义的方向。在这种背景下，本文对下一代网络中的资源共享环境进行了系统的研究，并取得了一系列有价值的研究成果。概括起来，本论文的创新成果主要体现在以下几个方面：(1)提出了一种融合网格的下一代网络资源共享模型—GRSM，并基于GRSM首次提出了融合网格的下一代网络资源共享环境—GRSEN，解决了下一代网络的资源共享问题。所提出的GRSEN作为GRSM的实现实例借助于网格技术将下一代网络资源统一封装成网格服务，并定义了虚拟服务和虚拟组织来解决NGN资源共享问题。采用离散数学和Z语言对GRSM的静态特性和动态特性进行了形式化定义，从而准确描述了GRSM解决下一代网络资源共享问题的方法。从资源共享的角度定义了GRSEN的资源模型，为GRSEN的资源共享、信息服务以及共享资源的管理奠定了基础。(2)提出了一种融合网格技术的下一代网络服务支撑环境-GOSSEN，并基于GOSSEN提出了NGN业务能力开放模型及业务交互模型，从而解决了下一代网络网络能力和业务能力以一致的方式开放的问题，进而为下一代网络的资源共享奠定了基础。在GOSSEN的核心组件—融合网格的增值服务执行环境(GSEE)的设计中，利用网格技术对资源共享和协同的支持，在系统过载时采用“疏导为主”的方法，将服务请求调度到有空闲处理能力的外部服务器中执行，从而解决传统NGN应用服务器处理能力不足的问题。(3)提出了一种新的基于软件框架的GOSSEN设计方法，该方法通过区别定义系统框架和构件框架，并对这两种框架分别采用黑盒框架设计和白盒框架设计的方法。提高了GOSSEN系统的可重用性，进而提高了系统开发效率和可靠性。(4)提出了GRSEN的基于优先级的任务调度SPN模型。分析GRSEN的具体要求，提出了改进的基于优先级的任务调度算法—PASED算法，该算法在满足基于优先级任务调度的基础上能有效提高系统吞吐量。所提出的GRSEN任务调度系统模型和算法对通用服务网格系统也具有普遍意义。更多还原

【Abstract】 The Next Generation Network (NGN) is the evolution goal of telecommunications network. With the development of network and the emergence of more and more service requirements, the realization of the resource sharing based on the opening of network capability and service capability becomes important. However, the research on the resource sharing in NGN remains inadequate. Grid technology is concerned with coordinated resource sharing and problem solving in dynamic, multi-institutional virtual organizations which can be used to solve the resource sharing problem in NGN. Within this background, this dissertation makes an in-depth and systematic research on resource sharing environment in NGN. And the main contributions are as follows:(1) This dissertation proposes a Grid-enabled Resource Sharing Model in NGN-GRSM. Based on GRSM, a Grid-enabled Resource Sharing Environment in NGN-GRSEN is also proposed. GRSM and GRSEN solve the resource sharing problem in NGN. GRSEN wrappes all the resources in NGN into Grid Service and defines the virtual organization and virtual service to solve the resource sharing problem with the help of Grid technology. The formalization method of GRSM and its service access mechanism is also proposed which describes the static and dynamic characterastic of the model precisely. The resource model of GRSEN is also proposed which is the basis of the resource sharing, information service and the management of GRSEN.(2) GOSSEN-a Grid-enabled Open Service Supporting Environment in NGN is proposed in this dissertation which is an important system of GRSEN. A Service Capability Opening Model and Service Interaction Model in GOSSEN are proposed to realize the opening of network capability and service capability in an uniform way which build the basic of resource sharing in NGN. The task scheduling system of GSEE (Grid-enabled value-added Service Execution Environment) which is the core component of GOSSEN is studied. With the help of Grid technology, GSEE can leverage all the spare resources to process service request, and then solve the performance bottleneck problem with traditional Application Server in NGN.(3) A software framework based design method of GOSSEN is proposed, which defines the System Framework and Component Framework independentely, and the black-box and white-box framework are applyed to the System Framework and Component Framework seperately. The proposed design method improves the reusability of GOSSEN and promotes the development effeciency and reliability of GOSSEN.(4) A SPN model of the priority-based task scheduling system of GRSEN is proposed. Based on this SPN model, the task scheduling requirements of GRSEN are analyzed and an advanced priority-based task scheduling algorithm (PASED) is proposed. The simulation results verify that the proposed algorithm can not only schedule the GRSEN task according to priorities, but also promote the throughput of GRSEN. The proposed SPN model and scheduling algorithm are meaningful to common service grid system.更多还原