QQ客服 官方微博 反馈留言

题名	普适计算环境下服务组合及部署的研究
作者	唐磊
学位类别	博士
答辩日期	2008-06-05
授予单位	中国科学院软件研究所
授予地点	软件研究所
关键词	普适计算 上下文 上下文协商 服务组合 服务选取 服务部署
其他题名	Research on Service Composition and Deployment for Pervasive Computing
中文摘要	普适计算（Pervasive Computing）的出现改变了传统的以计算机为中心的资源和信息获取模式，提出一种以人为本的计算方式，为人们提供随时随地、适应环境变化的资源和信息的获取方法，成为下一代网络的主流发展方向之一。普适计算具有环境异构并且动态变化的特点，在这种网络环境中信息的开发和集成打破了传统的模式，软件变服务的观点被普遍接受，并且得到了广泛的应用。因此，面向服务的计算（Service-oriented Computing）和普适计算的结合，是未来信息技术的下一阶段，也成为目前研究的热点问题。 一方面针对普适计算的特点，研究其动态变化、移动性和用户相关性等；一方面针对面向服务的计算，研究其体系结构、服务的描述、查询、组合、部署等问题。普适计算的特点决定了在该环境下的服务组合需要适应环境的动态变化；服务部署需要满足资源约束条件，提高网络资源利用率。基于研究背景以及对目前存在问题的分析，本文进行普适计算环境下服务组合及部署的研究，主要工作包括以下四个方面： （1）提出了一种规则驱动上下文相关的服务组合方法 服务组合是面向服务的计算中经典问题，研究成果包括组合模型、语言、方法和系统等，根据需求提出基于工作流、基于形式化方法、基于语义以及基于服务非功能属性等多种服务组合方法，但这些方法难以解决普适计算中的服务组合问题。因为在普适计算环境下，服务组合不仅需要满足功能性要求，而且要适应环境的异构和动态变化，能够根据环境变化自动的构建组合路径并动态的绑定服务。本文使用上下文刻画普适计算的动态变化性，提出一种基于本体的上下文模型和描述语言，并扩展服务描述，将上下文引入服务组合，提出服务组合算法，实现了规则驱动上下文相关的服务组合路径的建立。 （2）提出一种基于上下文协商的动态服务选取方法 服务组合路径建立了服务执行的先后顺序，在服务的执行期间需要动态的绑定实例，绑定服务实例过程称为动态的服务选取。考虑到普适计算中设备资源有限性以及提高网络资源利用率，本文提出一种基于上下文协商的服务选取方法。首先从已有的上下文本体中获取与动态服务选取相关的上下文，并根据其内容的不同分为用户上下文、设备上下文和服务上下文；然后给出上下文协商的约束条件，基于约束条件提出了上下文的协商算法，进行服务的选取绑定，实现动态服务组合。最后通过实验验证了算法的效率和有效性，分析了上下文数目和服务数目的变化，对服务组合成功率、资源利用率、服务选择冲突率等的影响。 （3）提出普适计算环境下的服务动态部署问题及算法 某些特定的应用（如数字家庭等）需要将服务部署在一个普适计算环境中，部署时需要考虑网络中负载均衡，提高资源利用率。本文针对普适计算中设备异构、资源有限的特点，给出组合服务动态部署问题的描述，并说明该问题是一个NP问题，提出一个启发式算法，能够在多项式时间内找到近似最优解，最后通过算法模拟和实验验证了算法的有效性以及时间性能，说明该方法适用于设备或者原子服务数目较多的普适计算环境。 （4）实现了一个服务组合原型系统 本文在国家八六三项目和国家科技攻关项目研究的基础上，实现了一个服务组合原型系统，该系统是“Web服务计算平台”的一部分，并以分布式文档管理系统为例，给出了服务组合的实现效果。
英文摘要	Pervasive computing has been put forward to go beyond the traditional acquirements of resources and information, which based on PCs. It provides a human-centered computing pattern with information and communication technology everywhere, for everyone, at all times and becomes a mainstream of the next generation network. The main features of pervasive computing environment are ubiquitous, heterogeneous and dynamic. In this environment, the developing and integrating of information are different from the old methods, and the viewpoint that a software is a service has been accepted widely. Therefore, the integration of service-oriented computing with pervasive computing has become an active area of research. For pervasive computing, the research work includes the dynamic environment, the mobile device, user-involved, etc. For service-oriented computing, the research work includes the architecture, service description language, service finding, composition, selection, deployment, etc. This thesis puts forward an approach to service composition and deployment in pervasive computing environment. The main points of this thesis are described as follows: (1) The thesis provides an approach to ECA rules driven context-aware service composition Service composition is an important problem in service-oriented computing. The research efforts include service composition model, language, framework, etc. Many service composition methods, such as workflow-based, semantic-based, QoS-aware or context-aware, are discussed for different requirements and applications. However, in pervasive computing, service composition should not only satisfy the functional requirements but also be adapt to the heterogeneous and dynamic environment. The thesis uses context to describe the pervasive computing environment, and provides an ontology-based context modeling language and context reasoning. Based on the context description, the web ontology language for service has been extended with context. It also puts forward EAC rule driven context-aware service composition. (2) An approach to service selection based on context negotiation has been put forward Before a service instance is invoked and executed, it should be selected and bound to adapt to the pervasive computing and satisfy the user’s preferences. First, the definition of the context negotiation is given. Based on the definition, the thesis gives the constraint conditions of context negotiation. According to the constraint conditions, it presents the service composition algorithm. Furthermore, the algorithm is optimized to improve the resource utilization. Some experiments are given to validate the algorithms performance and validity. The experiment data prove that the approach presented in the thesis can be applied to context sensitive service composition in pervasive computing. (3) It discusses the service dynamic deployment method in pervasive computing Based on Liquid—an embedded and component-based system, this thesis presents a dynamic deployment problem and the algorithm of composite services for pervasive computing. First, the thesis defines the problem of dynamic deployment in the pervasive computing environment. Second, the randomized algorithm and the heuristic algorithm are given to solve the problem. Finally, the experiment data are given to analysis and compare the performance of the different algorithms. According to the method, the composite services are deployed into various embedded devices to satisfy resource requirement and improve the resource utilization. The algorithm simulation and analysis indicate that it can be applied to the pervasive computing environment with more devices or atomic services. (4) The service composition prototype is designed and implemented Based on the National High-tech Research and Development Plan of China and National Key Technologies Research and Development program, the thesis designs and implements a prototype of context-aware service composition and applies it to document management system for pervasive computing. It solves the problem that an organization manages the documents distributed in network concentratively.
语种	中文
公开日期	2011-03-17
页码	118
内容类型	学位论文
源URL	[http://124.16.136.157/handle/311060/5798]
专题	软件研究所_基础软件国家工程研究中心_学位论文
推荐引用方式 GB/T 7714	唐磊. 普适计算环境下服务组合及部署的研究[D]. 软件研究所. 中国科学院软件研究所. 2008.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。