软件定义存储模型的研究及实现

发布时间：2018-07-08 10:18

本文选题：软件定义存储 + 分布式文件系统　；参考：《华北电力大学(北京)》2017年硕士论文

【摘要】：伴随着云计算技术的广泛应用,互联网及物联网技术的不断快速发展,大数据时代信息数据量出现了爆发性增长的态势,传统架构下的存储系统已经难以满足不同行业的信息系统在兼容性、统一管理、扩展性、安全性、差异化服务等方面的需求。在云计算应用场景下,海量的数据被集中存储在一起,随着数据量的飞速递增以及存储规模的不断扩大,数据读写性能瓶颈变得十分严重,因此对存储系统的I/O性能、总体容量、可扩展性和可管理性等方面提出更高的要求。因为存储厂商在各自存储设备的实现方式上长期存在技术差异和严重的技术壁垒,传统架构存储系统大多是单一、封闭、集成的系统,只支持特定软件和硬件的组合,存储系统缺乏灵活性,无法充分利用不断涌现的新硬件产品的能力和新平台,并且只能进行有限度的扩展,无法满足用户快速增长的数据存储的需要。同时,不同存储系统之间只能实现非常有限的统一监控,而无法形成完整统一、有效调度、全局管理的存储资源池。当今许多企业的数据中心都在经历从传统数据中心到云架构的转型,在转型过程中,存储架构成了关键的环节。软件定义存储从提出以来,就被广泛认为是下一代存储技术发展方向,软件定义存储是将硬件驱动从存储技术中剥离出来,通过软件定义反向驱动硬件,将与存储相关的控制工作都放置在相对于物理存储硬件的外部软件中,可以提供包括存储虚拟化,存储资源动态管理等功能。用户可以利用软件定义技术,来打破厂商之间的限制壁垒,消除专用平台限制,建设可灵活扩展低成本的面向云计算的存储系统。本文软件定义存储的模型研究旨在利用新的软件定义存储技术设计一套适用于电力系统的软件定义存储模型,通过软件定义的方式实现存储系统软硬件的解耦,使其具备灵活的配置策略、多元化的存储供给能力以及存储资源的高效管理。
[Abstract]:With the wide application of cloud computing technology and the rapid development of Internet of things and Internet of things, the amount of information data in the era of big data has increased dramatically. The traditional storage system has been difficult to meet the needs of different industries in terms of compatibility, unified management, scalability, security, differentiated services and so on. In the cloud computing application scenario, the massive data is stored together. With the rapid increase of the amount of data and the expansion of the storage scale, the bottleneck of data read and write performance becomes very serious, so the I / O performance of the storage system becomes very serious. Overall capacity, scalability and manageability are higher requirements. Because storage manufacturers have long had technical differences and serious technical barriers in the implementation of their respective storage devices, most of the traditional architecture storage systems are single, closed and integrated systems, which only support the combination of specific software and hardware. The storage system lacks flexibility, can not make full use of the new hardware products and new platform, and can only be limited expansion, can not meet the rapid growth of user data storage needs. At the same time, different storage systems can only achieve a very limited unified monitoring, but can not form a complete unified, effective scheduling, global management of storage resources pool. Nowadays, many enterprise data centers are undergoing the transition from traditional data centers to cloud architectures. In the process of transformation, storage architecture has become a key link. Since software definition storage was put forward, it has been widely regarded as the development direction of the next generation storage technology. Software definition storage is to separate the hardware driver from the storage technology and to define the reverse drive hardware through software. The control work related to storage is placed in the external software relative to the physical storage hardware, which can provide functions such as storage virtualization, dynamic management of storage resources, and so on. Users can use software definition technology to break down barriers between vendors, eliminate restrictions on dedicated platforms, and build a flexible, low-cost storage system for cloud computing. The purpose of this paper is to design a set of software definition storage model for power system by using new software definition storage technology, and to decouple the software and hardware of storage system by the way of software definition. It has flexible configuration strategy, diversified storage supply capacity and efficient management of storage resources.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP333

【参考文献】