多媒体服务覆盖网中的服务组合问题研究

发布时间：2018-06-06 07:11

本文选题：多媒体服务 + QoS　；参考：《东北大学》2008年博士论文

【摘要】： 随着Internet上Web服务的快速发展,越来越多的服务提供商以Web服务的形式为用户提供多媒体应用。但由于上述应用功能单一,服务能力有限,如何利用服务组合技术向用户提供综合、复杂的多媒体服务成为一个有实际应用价值的课题。为此,本文针对分布式多媒体应用中的服务组合问题,利用web服务间能够互联形成一个应用层服务覆盖网的特点,提出了基于多媒体服务覆盖网的服务组合框架,并围绕该框架中的服务发现、服务路由以及服务路径的失败恢复等问题展开研究,通过服务组合为用户提供智能的集成服务,满足用户动态、多变的功能及非功能性需求。主要工作如下： (1)提出了一个基于多媒体服务覆盖网的服务组合框架,满足Internet上大规模分布式多媒体应用需求。针对多媒体应用不能保证服务质量(QoS),缺乏灵活性和可扩展性等问题,提出了一个基于多媒体服务覆盖网的分层服务组合框架MSON。MSON从服务组合模式和多媒体服务组合模型出发,采用层次结构来简化复杂的服务组合问题,分离出其基本功能单元,建立层次化结构模型,为分布式多媒体应用提供一致的系统框架,实现应用的可扩展性和灵活性。 (2)提出了一个分布式语义服务发现方法,解决MSON中服务发现问题。针对传统UDDI服务发现方法不能解决MSON面临的分布式查找和语义服务匹配问题,对MSON提出了相应的一个分布式服务发现方法。该方法在服务注册节点上发布语义服务广告,采用Bloom filter技术提高搜索效率用预计算方法回避高耗时的推理操作,提高服务匹配效率。针对从众多功能相似的服务中选择最优服务问题,从信任和信誉管理的角度提出基于QoS的服务选择排序算法。该算法根据用户的反馈、服务提供者提供的数据以及可信代理的报告三方面的因素计算QoS预期值,并根据QoS预期值对服务选择和排序。 (3)提出了QoS服务路由算法,解决MSON中服务路径选择问题。MSON是通过把多媒体应用需求映射为覆盖网路由的方式实现服务组合的,本文从平面覆盖拓扑和分层覆盖拓扑两个角度研究服务路由算法,提出满足QoS的平面服务路由算法KTA-LBSPSA和QoS感知的层次服务路由算法HQSRA. KTA-LBSPSA算法首先在建立的最小延迟服务树中分层查找满足QoS约束的服务路径,并从多个候选服务路径中选择有最大负载公平值的路径作为最终的服务路径,得到一个满足QoS约束、负载均衡的服务路径。HQSRA算法将服务路径选择过程分为簇间和簇内选择两个步骤,解决在大规模分层覆盖网中发现QoS满意的服务路径问题。 (4)提出了一个最小干扰服务组合和恢复算法,解决MSON中服务路径稳定性问题。针对MSON覆盖网中的服务节点和链路存在失效的问题,提出一个优化的服务干扰模型来定量地描述服务中断对端用户的影响；并遵循优先使用可本地恢复的节点、选择可靠节点以及消极调整服务路径三个路由选择原则,设计了最小干扰服务组合和恢复启发式算法MISCR.仿真实验表明,相比传统的算法,MISCR降低了服务中断对端用户的影响,提高了系统的健壮性和稳定性。
[Abstract]:With the rapid development of Web services on Internet, more and more service providers provide multimedia applications for users in the form of Web services. However, because of the single application function and limited service ability, how to use service composition technology to provide comprehensive services to users, complex multimedia services have become a practical application value topic. This paper, aiming at the problem of service composition in distributed multimedia applications, makes use of the characteristics of Web Services interconnected to form an application layer service coverage network, and puts forward a framework of service composition based on multimedia service coverage network, and focuses on service discovery, service routing and service path failure recovery in this framework. The research provides users with intelligent integrated services through service composition to meet users' dynamic, changeable and non functional requirements. The main tasks are as follows:
(1) a service composition framework based on multimedia service coverage network is proposed to meet the needs of large-scale distributed multimedia applications on Internet. In view of the problem that multimedia applications can not guarantee the quality of service (QoS), lack of flexibility and extensibility, a hierarchical service composition framework based on multimedia service coverage network (MSON.MSON) is proposed. From the service composition model and the multimedia service composition model, the hierarchical structure is used to simplify the complex service composition problem. The basic functional units are separated and the hierarchical structure model is established. It provides a consistent system framework for distributed multimedia applications, and realizes the extensibility and flexibility of the application.
(2) a distributed semantic service discovery method is proposed to solve the problem of service discovery in MSON. In view of the problem that the traditional UDDI service discovery method can not solve the problem of distributed lookup and semantic service matching faced by MSON, a corresponding distributed service discovery method is proposed for MSON. This method releases semantic services on service registration nodes. Bloom filter technology is used to improve the search efficiency by precomputing the time-consuming reasoning operation and improving the service matching efficiency. A service selection sorting algorithm based on QoS is proposed from the perspective of trust and reputation management. The algorithm is based on the user's feedback and service. The data provided by the provider and three factors of the trusted agent's report calculate the expected value of QoS, and select and sort the services according to the expected value of QoS.
(3) the QoS service routing algorithm is proposed to solve the service path selection problem in MSON..MSON is a service combination by mapping the multimedia application requirements to the overlay routing. This paper studies the service routing algorithm from the two angles of the flat cover topology and the layered overlay topology, and proposes a QoS based plane service routing algorithm KTA-LBSP. The SA and QoS aware hierarchical service routing algorithm HQSRA. KTA-LBSPSA algorithm first delaminate the service path to satisfy the QoS constraint in the minimum delay service tree, and select the path of the maximum load fair value from multiple candidate service paths as the final service path, and get a service that satisfies the QoS constraint and load balance. The path.HQSRA algorithm divides the service path selection process into two steps between cluster and cluster, and solves the problem of finding QoS satisfied service path in the large scale layer coverage network.
(4) a minimum interference service combination and recovery algorithm is proposed to solve the problem of service path stability in MSON. In view of the problem of service nodes and links in the MSON overlay network, an optimized service interference model is proposed to quantitatively describe the effect of service interruption on the end users and to follow the priority use and can be recovered locally. Nodes, selected reliable nodes and negative adjustment of service path three routing principles, designed the minimum interference service composition and the recovery heuristic algorithm MISCR. simulation experiments show that compared with the traditional algorithm, MISCR reduces the impact of service interruption on end users, and improves the robustness and stability of the system.
【学位授予单位】：东北大学
【学位级别】：博士
【学位授予年份】：2008
【分类号】：TN919.8

【相似文献】