【作者】 曾文英；

【导师】 赵跃龙；

【作者基本信息】 华南理工大学 ， 计算机应用技术， 2011， 博士

【摘要】 随着移动网络技术的快速发展和移动用户数量的持续增长,移动环境下的各种业务应用已经日趋广泛。伴随而来的是移动应用环境下产生的数据信息呈指数式的增长,它带来了对移动环境下的大量数据信息进行存储管理的新技术需求,因此,有关面向移动环境下的数据存储管理问题的研究已经变得越来越重要。现有的有线网络中的分布式存储管理方法主要是针对带宽稳定、可持续服务、可扩展和高性能节点等都比较稳定的应用环境;但它在面对移动环境的异构性、分布性、高维性、动态性带来的移动数据管理的复杂性等方面具有明显的不适应性,因此,不能直接应用于移动环境下的数据存储管理。由于移动环境的多源性、多宿性、自治性、上下文感知性和环境依存性,且面向移动环境的数据存储管理具有集中与分布的特点,因此,可以考虑借助移动Agent技术、移动数据库、分布式网络、跨层协作、网格计算及云计算等多种技术来构建移动环境存储管理与服务系统。另外,随着各种网络互通和融合技术的日益成熟,在移动环境下设备与网格和云计算系统协作可以构建海量、持久、无限可扩展的存储资源与服务系统的市场前景巨大,并且良好的数据存储管理方法将会为移动应用的运行服务提供高效、安全的数据存储基础。因此,面向移动环境的数据存储管理方法的研究有重要的理论意义和实际意义。本文对移动环境下的移动网格体系结构与资源选择方法、移动分布式数据存储服务结构模型、移动数据库技术、基于无线Mesh网的层次化存储系统和移动环境下的存储服务QoS等关键技术问题进行了系统而深入的研究,取得了一些有创新性意义的研究成果;其主要研究工作和创新性成果体现在以下几个方面。1.提出了一种移动网格结构模型(MGAM,Mobile Grid Architecture Model)和移动网格资源管理算法(MGRMA,Mobile Grid Resource Management Algorithms)。首先提出了一种结合移动计算和网格计算、支持移动环境存储服务的移动网格结构模型,其次,对移动网格的逻辑构成和形式化模型进行分析,给出了移动网格资源选择与分配方法和相关的移动网格资源协作算法,最后给出了移动网格的原型和应用实例,并进行了模拟测试和性能分析。2.提出了一种基于移动环境的自适应分布式存储服务的系统结构模型(SDSSAM,Self-Adaptive Distributed Storage Service Architecture Model),SDSSAM是一种结合移动计算的跨层协作式存储结构模型。首先描述了SDSSAM的各层次的功能;其次提出了SDSSAM结构中的跨层协作方式;最后给出了SDSSAM的分布式存储协作服务和自适应存储等算法。研究表明SDSSAM具有自适应、移动计算、分布式计算和自组织的特点,是一种具备了灵活性、自主性、协作性和群体智能的移动存储系统结构。3.提出了一种基于移动数据库的移动数据管理结构(MDMA,Mobile Data Management Architecture)和存储管理方法(SMS,Storage Management Solution)。移动数据库是移动分布式环境数据组织和存储的最有效的方式,为移动业务运行提供了数据支撑,移动应用一般基于移动数据库而实现。针对移动环境的特点,首先提出了一种基于移动数据库的移动数据管理结构和存储管理方法;其次研究了移动数据库中数据的预取与复制、缓存同步、事务处理、并发控制、广播机制等多种关键技术,为移动环境下数据存储与管理相结合提供可行的管理方法。4.提出了一种基于无线Mesh网的层次化存储系统(HSSWMN,Hierarchical Storage System over Wireless Mesh Network)模型。首先提出了基于无线Mesh网的层次存储系统(HSSWMN)模型,并对其存储模型、存取算法、性能优化等方面进行了分析和研究;其次研究了HSSWMN的名字空间与元数据服务、搜索与查找服务、注册与注销,可扩展性、负载均衡、容错机制、数据安全、复制与缓存机制和拓扑重构等关键问题;最后通过仿真分析,对时延、吞吐量、误码率等进行了模拟测试,并对HSSWMN存储系统的可行性、可用性和可靠性进行了性能分析。5.提出了一种面向移动环境数据存储服务QoS跨层模型(QCLMSS,QoS Cross Layer Model of Storage Services)和移动环境存储服务QoS确保算法(QASS,QoS Guarantee Algorithms of Storage Service)。首先对移动环境数据存储服务QoS技术进行了研究,分析了各层次QoS的特征及关系;其次提出了QoS实施算法及性能模型,并对移动环境下存储QoS保证算法进行了研究;最后提出了全局优化、局部优化、多阶段优化、自适应优化等算法,并分析了移动存储系统的QoS实例,对有线网络、无线网络接入方式下磁盘I/O性能进行了模拟测试和分析研究。更多还原

【Abstract】 Along with the rapid development of mobile network and communication technology, and continuous growth of mobile users’number, various business applications in mobile environment have become widely and pervasively. Consequently, the data and information show exponential increasement under mobile applications environment, which bring about the new technology requirements to store and management the massive data and information. For these reasons, the mobile environment oriented data storage and management questions will become more and more important.Nowadays distributed management methods of wired network mainly points for the stable environment with width stability, continuous services, and scalability and high performance nodes. The methods will not be adaptable for the mobile data management complexity of mobile environments with heterogeneity, distributed, high dimensions, dynamicity and movability, and so they doesn’t be directly applied into data management in mobile environment. For the multi-sources, multi-sinks, self-control, context aware and environment dependability, and the centralization and distribution coexistence of mobile environment oriented data storage management, there are many technologies which may be combined with to construct mobile environment storage management and service system by mobile agent, mobile database, distributed network, cross layer cooperation, grid computing, cloud computing and etc. Furthermore, along with the internetwork and convergence technology becoming adult, mobile environmental devices may be cooperated with grid computing and cloud computing system to construct massive, persistent, boundless scalable storage resources and service system, which will bring about huge market prospect, and good data storage management methods will provide high efficient, secure data storage foundation. Consequently, mobile environment oriented data storage management methods will show important theoretical and practical meanings.The dissertation makes systematical and deepen researches on the key technologies of mobile grid architecture and resource selection, mobile environment oriented distributed data storage service structure model, mobile database technology, hierarchical storage system based on wireless mesh network, QoS of storage service under mobile environment and etc., and makes several creative research achievements, and the main research work and creative achievements of the dissertation lie in the following aspects.1. A mobile grid architecture model of mobile environment (MGAM, Mobile Grid Architecture Model) and mobile grid resources management algorithm (MGRMA, Mobile Grid Resource Management Algorithm) are proposed. Firstly, the dissertation presents a mobile grid architecture model combined with mobile computing and grid computing, supported mobile environment storage services. Secondly, logic components and formal model of mobile grid are analyzed and mobile grid resource selection and assignment method and mobile grid resources cooperation algorithms are proposed. At last, the prototype of the mobile grid and application examples are given, including mobile grid prototype and performance analyses and simulation test.2. A self-adaptive distributed storage service architecture based on mobile environment (SDSSAM, Self-Adaptive Distributed Storage Service Architecture Model) is proposed. SDSSAM is a cross-layer collaborative model combined with mobile computing. At first, every layer function of SDSSAM is described. Second, collaboration mode of cross-layer is illustrated. At last, distributed storage cooperation service and self-adaptive storage algorithm on SDSSAM are given out. Research shows that SDSSAM has self-adaptive, mobile computing, distributed computing and self-organized characteristics, and it is a kind of mobile storage system architecture with advance flexibility, self-control, collaborative and swarm intelligence.3. Mobile data management architecture based on mobile database (MDMA, Mobile Data Management Architecture) and storage management solution (SMS, Storage Management Solution) are proposed. Mobile database is the most efficient mode for mobile distributed environment data organization and storage, which will provide data foundation of mobile business operations, because mobile applications are implemented based on mobile databases. Aiming at the characteristics of mobile environment, mobile data management architecture and storage management solution are proposed, and the pre-fetch and replication, cache synchronization, transaction process, concurrence control, broadcast mechanism and other key technologies are discussed. These will provide available and conventional runtime methods for the data storage and management under mobile environment.4. A model of hierarchical storage system based on wireless mesh network (HSSWMN, Hierarchical Storage System over Wireless Mesh Network) is proposed. Firstly, the dissertation proposes the hierarchical storage system model based on wireless mesh network (HSSWMN), and the storage model, access algorithms and performance optimization methods are analyzed and researched. Secondly, name space and metadata service, search and looking for service, registration and logout, scalability, load balance, fault tolerant, data safety, replication and cache mechanism, topology reconstruction of HSSWMN are researched. At last, by the simulation analysis, simulation testing discloses that the delay, throughput, error code and etc, performance analysis is carried for the feasibility, availability and reliability of HSSWMN storage system.5. A QoS cross layer model of storage service of mobile environment oriented (QCLMSS, QoS Cross Layer Model of Storage Services), and QoS guarantee algorithm of storage service (QASS, QoS Guarantee Algorithms of Storage Service) are proposed. The dissertation first does research on the QoS technologies of mobile environment storage service, analyzes the properties and relations between layers. Secondly, the QoS guarantee algorithms and performance model are proposed, and researches are done on QoS guarantee algorithms. At last, Global optimization, local optimization, mluti-state optimization, and self-adaptive optimization algorithms are proposed, and analysis and comparation of the mobile storage system QoS examples are analyzed, simulation test and analysis research is taken on disk I/O performance of wired network and wireless network access mode.更多还原