应用于OpenStack平台的无元数据服务器的海量网络存储系统设计

发布时间：2018-04-09 19:48

本文选题：云存储　切入点：OpenStack平台　出处：《浙江大学》2014年硕士论文

【摘要】：当前,随着时代的发展与数据爆炸性的增长,传统的计算模式已经越来越不能适应时代发展的需要,取而代之的是云计算和大数据。而云存储可以视为云计算的一部分,这个概念是在云计算概念的基础上衍生和发展出来的一个新的概念,云存储系统可以视为一个以数据存储和管理为核心的云计算系统。 OpenStack是一个由美国国家航天局和Rackspace合作研发的云平台管理项目,可以为普通用户建立和部署云计算服务。其具有强大的功能,但也存在着不足。本文正是基于此种背景,设计了一种无元数据服务器的海量网络存储系统。本文首先介绍了OpenStack平台和网络存储系统,然后对无元数据服务器的网络存储系统的进行了设计,包括存储系统的架构,检索算法的设计,可靠性设计和读写策略等,最后将无元数据服务器的海量网络存储系统与OpenStack进行了对接,使其可以为OpenStack提供存储。本文的重点是无元数据服务器的海量网络存储系统设计。传统的网络存储系统都具有元数据服务器,限制了整个系统的性能,产生了性能瓶颈和单点故障问题。本文创新性地在海量网络存储系统中取消了元数据服务器,改为采用算法实现存储系统中文件的检索、访问和调度,消除了性能瓶颈和单点故障问题。本文的另一个创新点是采用了多元环的角度来看待网络存储系统中的映射,并提出了一种快速计算多元环的方法,该方法比以前的方法更有效率。 OpenStack默认只支持块存储设备,这很大程度上限制了其应用。本文创新性地将所设计的无元数据服务器的海量网路存储系统应用于OpenStack平台,使其突破了块存储的限制,存储功能更加强大。
[Abstract]:At present, with the development of the times and the explosive growth of data, the traditional computing mode has become more and more unable to meet the needs of the development of the times, instead of cloud computing and big data.Cloud storage can be regarded as a part of cloud computing. This concept is a new concept derived from and developed from the concept of cloud computing. Cloud storage system can be regarded as a cloud computing system with data storage and management as its core.OpenStack is a cloud platform management project developed by NASA and Rackspace to build and deploy cloud computing services for ordinary users.It has the powerful function, but also has the insufficiency.Based on this background, this paper designs a massive network storage system without metadata server.This paper first introduces the OpenStack platform and the network storage system, then designs the network storage system of the non-metadata server, including the architecture of the storage system, the design of the retrieval algorithm, the reliability design and the strategy of reading and writing, etc.Finally, the massive network storage system of the non-metadata server is docked with OpenStack so that it can provide storage for OpenStack.The key point of this paper is the design of massive network storage system of non-metadata server.Traditional network storage systems have metadata servers, which limit the performance of the whole system, and cause performance bottlenecks and single point failures.This paper innovatively cancels the metadata server in the massive network storage system, and instead uses the algorithm to realize the retrieval, access and scheduling of the files in the storage system, which eliminates the performance bottleneck and the single point fault problem.Another innovation of this paper is that the mapping in network storage system is viewed from the point of view of multivariate ring, and a fast method to compute multivariate ring is proposed, which is more efficient than the previous method.OpenStack only supports block storage devices by default, which greatly limits its application.This paper innovatively applies the massive network storage system of the non-metadata server to the OpenStack platform, which breaks through the limitation of block storage, and the storage function is more powerful.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP333

【相似文献】