基于KVM云计算平台的分布式关系型数据库的设计与实现

发布时间：2018-11-26 06:59

【摘要】：自计算机诞生以来,数据的存储和处理便是计算机信息技术领域的重要部分,在20世纪60年代就已经提出了数据库管理系统概念,E.F.Codd在其论文中阐述了关系模型,关系型数据库逐渐成为主流。随着信息技术和软件工程的发展和大数据时代的到来,传统的关系型数据库虽然有其使用上的便利性,但在性能上存在瓶颈,难以直接应对现在的数据规模和时效性要求,因此将传统数据库单点架构转换为分布式架构成为一种有效的解决方法,各个不同的厂商或机构都对这一领域进行了不同的探索,也提出了一些不同的解决方案和模型。在实际部署或使用过程中,一套完整的数据分析处理系统或信息服务系统模型被细分为更多的层次模块。往往在搭建过程中,需要负责不同部分的专业人员协同工作,增加建设成本。因此,如何能保留传统关系型数据库低使用门槛的优点,同时又能保证其在处理大规模数据的性能具有十分重要的研究意义。与此同时,云计算由于其特有的资源弹性伸缩、集中管理、数据安全等优点开始被广泛使用,用户可以根据自己的需要购买相应资源而且在短时间内便可直接使用,免去了自行搭建硬件平台和网络的工作,同时也省去了自行维护的任务,更多的信息服务商开始将自己的平台搭建在云计算平台之上。因此,本论文所设计的系统,也希望能够充分利用云平台的资源灵活分配的特性。综上,本文针对在将关系型数据库转为分布式架构的过程中,数据库功能和性能上面临的主要技术难题,例如分布式事务处理,数据快速定位等等问题,进行了研究。本论文的主要研究目的是提供一种基于云计算平台的分布式数据库,通过探究相应的数据分布策略、数据查询使用模型、数据自动同步方法以及系统弹性伸缩模型,通过合理的模型设计实现一种低成本、易使用的、弹性可伸缩的分布式关系型数据库系统。论文首先对分布式数据库的课题背景和国内外研究现状做了综述,并交代了论文结构安排,然后提出了在系统模型设计上所存在的重点难点问题,同时根据作者的相关技术经验,提出了相应的解决方法和整体的系统设计方式,最后进行了部分功能和性能的测试以及总结分析。
[Abstract]:Since the birth of computer, the storage and processing of data is an important part of computer information technology. In the 1960s, the concept of database management system has been put forward, and E.F.Codd has elaborated the relational model in his thesis. Relational database has gradually become the mainstream. With the development of information technology and software engineering and the arrival of big data era, although the traditional relational database has its convenience to use, there are bottlenecks in its performance, so it is difficult to directly deal with the demand of current data scale and timeliness. Therefore, the conversion of traditional database single point architecture to distributed architecture becomes an effective solution. Different manufacturers or organizations have made different explorations in this field, and have also put forward some different solutions and models. In the actual deployment or use process, a complete data analysis and processing system or information service system model is subdivided into more hierarchical modules. Often in the process of building, need to take charge of different parts of the professionals to work together to increase the cost of construction. Therefore, it is of great significance to study how to retain the advantages of low threshold of use in traditional relational databases, and at the same time to ensure their performance in dealing with large scale data. At the same time, due to its unique advantages of flexibility, centralized management, data security and so on, cloud computing has been widely used. Users can purchase the corresponding resources according to their own needs and can use them directly in a short period of time. It eliminates the work of building hardware platform and network, and also saves the task of self-maintenance. More information service providers begin to build their own platform on cloud computing platform. Therefore, the system designed in this paper also hopes to make full use of the flexible resource allocation of cloud platform. In summary, this paper focuses on the main technical problems in the process of transforming relational database into distributed architecture, such as distributed transaction processing, fast positioning of data, and so on. The main purpose of this paper is to provide a distributed database based on cloud computing platform, through exploring the corresponding data distribution strategy, data query usage model, data automatic synchronization method and system elastic expansion model. A low cost, easy to use, flexible and scalable distributed relational database system is implemented by reasonable model design. In this paper, the background of the distributed database and the current research situation at home and abroad are summarized, and the structure of the paper is described. Then, the key and difficult problems in the design of the system model are put forward. At the same time, according to the author's relevant technical experience, the corresponding solution and the whole system design method are put forward. Finally, some function and performance tests and summary analysis are carried out.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【参考文献】