重庆师范大学录取线云存储服务系统研究外文文献翻译

2020年12月31日发(作者：江衡)
毕业设计附件
外文文献翻译：原文+译文
文献出处： Mehra P. Cloud computing: Distributed internet computing for IT and
scientific research[J]. Internet Computing, IEEE, 2016, 1(5): 10-19.
原文
The study of cloud storage service system
Mehra P
Abstract
Cloud storage is a new concept, which developments and extensions in a cloud
computing, so to understand cloud storage is the first to know about cloud computing.
Cloud computing is a kind of super computing model based on Internet, in a remote
data center, tens of thousands of computer and server connected to a computer cloud.
Therefore, cloud computing allows you to experience even 10 trillion times a second
operation ability, have such a powerful computing ability to simulate nuclear
explosion, forecasting and market development trend of climate change. User through
a computer, laptop, mobile phone access to the data center, operation according to
their needs. With the acceleration development of the concept of cloud computing,
people began to looking for a new place for huge amounts of information in cloud
storage. The cloud (cloud storage) emerged from a widely attention and support.
Similar to the concept of cloud storage and cloud computing, it refers to the
application through the cluster, grid technology or distributed file systems, and other
functions, the network of a large number of various types of storage devices set up by
applying the software to work together, common external provide access to data
storage and business functions of a system.
Keywords: cloud storage, cloud storage service system, the HDFS
1 Introduction
The rise of cloud makes the whole IT industry in a significant change, from
equipmentapplication centered toward centered on information and this change will
cause a series of changes, and affect the technical and business mode two levels. The
biggest characteristic of the cloud is a mass, high performancehigh traffic and low
cost, and the biggest change is that its bring providers from sales tools to gradually
according to the actual use of tools to collect fees, from selling products to selling
services. Therefore, it can be said that cloud storage is not stored, but service. Cloud
storage but also has the following characteristics: strong extensibility, should not be
limited by the specific geographic location, based on the business component,
according to the use of fees, and across different applications. The research content of
this article for the study of the cloud storage service system based on HDFS, aims to
build a cloud storage service system based on HDFS, solve the enterprise mass data
storage problem, reduce the cost of implementing the distributed file system, promote
the Hadoop technology promotion. Cloud storage is widely discussed in the present
on the cloud computing concept of extension and development, to a large number of
different types of storage devices in the network integration, thereby providing access
to data storage and business functions. Hadoop distributed file system (HDFS) is the
underlying implementation of open source cloud computing software platform
Hadoop framework part, has the characteristic such as high transmission rate, high
fault tolerance, can be in the form of a flow to access the data in the file system, so as
to solve the access speed and security issues, achieve huge amounts of data storage
management.
2 Each big cloud storage products of the company
2.1 The Amazon’s strategy
Amazon is among the first to launch the cloud storage service enterprises.
Amazon first launch a service of cloud computing is Amazon web services (Amazon
web services, the AWS), the cloud computing service is composed of four core
components: simple arrangement, simple storage service, elastic computing cloud and
is still in the backs of the test. In August 2008, Amazon in order to enhance its efforts
on the cloud storage strategy, its Internet services add
compute cloud (ECZ).The vendor launched Elastic Block Storage (Elastic Block
Storage, EBS) products, and claims that the product can through the Internet service
form at the same time provide Storage and computing functions.
2.2 Nirvana and CDNetworks strategy
Famous cloud storage platform providers Nirvanix and content delivery network
service provider CDNetworks released a new cooperation, and strategic partnerships,
to provide the industry the only cloud storage and content delivery service integration
platform. Use it’s located in all parts of the world 63 content distribution node, the
user can store unlimited data on the Internet, and get good data protection and data
security guarantee. Cooperation will bring CDNetworks in cloud storage and Nirvanix
the same capacity, not only can safely store huge amounts of media content, and can
rely on CDNetworks data center to deliver data anywhere in the world in real time,
the two companies, said based on this partnership of cooperation, make it have better
overall media delivery ability, also helps users save 80% 90% of the construction of
its own storage infrastructure costs.
2.3 Google's strategy
The company in this year's FO developer technical conference announced called

Look from the function design, Google storage will refer to the Amazon s3, for
existing s3 user’s switch to Google storage service. Google storage services will
include RESTAPI agreement, to allow developers to download via Google account
provides authentication, data backup services. In addition, Google will also to outside
developers to provide network user interface and data management tools.
2.4 EMC’s strategy
EMC's cloud storage infrastructure solution is a kind of management system
based on strategy, the service provided can create different types of cloud storage
ability, for example, it can be for not paying customers to create file two copies, and
stored in different locations around the world, and for paying customers to create a
backup storage on October 5, and provides its all over the world access to the file of
higher reliability and faster access. In software systems, Atm0S including data
services, such as copying, data compression, data reduplication, with cheap standard
x86 server to hundreds of terabytes of hard disk storage has promised that
it automatically configure the new storage and adaptive ability of a hardware failure,
also allows the user to use W b manage service agreement and read. At present there
are three versions, Atm0S system capacity is respectively 120 TB otb, 24, and 36 orb,
All of them are based on x86 servers and support gigabit or 10 gb Ethernet
connection.
3 Cluster development of storage technology
The rise of cloud storage is upending the existing network storage architecture.
Facing the current pet bytes of mass storage requirements, the traditional SAN or
NAS will exist in the expansion of the capacity and performance bottlenecks. Such as
by its physical elements (such as the number of disk drives, the connected server
performance and memory size and the number of controller), can cause a lot of
functional limitations (such as: the number of file system support, snapshot or copy
number, etc.).Once encountered the bottleneck of storage system, it will constantly
encourage users to upgrade to a bigger storage system and add more management
tools, thus increasing the cost. Cloud storage the service mode of the new storage
architecture is demanded to keep very low cost, and some existing high-end storage
devices are obviously cannot meet this need. From the perspective of the practice of
Google company, they are not used in the existing cloud computing environment SAN
architecture, but use, is a scalable distributed file system GFS) this is a highly
efficient cluster storage is a scalable distributed file system, used in
large, distributed, on a visit to a large amount of data applications. It runs on ordinary
PC, but can provide strong fault tolerance, can give a large number of users with the
overall performance of the service. Cloud storage 130] is not stored, but service. Wan
and the Internet like a cloud, the cloud storage for users, not referring to a particular
device, but is a by many a collection of storage devices and servers. Users use the
cloud storage, not using a storage device, but use, is the entire cloud storage system
with a data access service. So strictly speaking, the cloud is not stored, but a service.
Cloud storage is the core of application software combined with a storage device, by
applying the software to realize the change of the service to the storage device.
4 Cloud storage system analyses
Compared with the traditional storage devices, cloud storage is not only
hardware, but a network devices, storage devices, servers, applications, public access
interface, access, and the client program such as a complex system composed of
multiple parts. Parts storage device as the core, through the application software to
provide access to data storage and business services. The structure of cloud storage
system model consists of four layers.
(1) The storage layer
Storage layer is the most basic part of the cloud storage. Storage devices can be a
fiber channel storage devices, or other storage devices. Storage devices are often large
number of cloud storage and distribution of many different areas, between each other
through the wide area network, Internet or fiber channel network together. Storage
devices is a unified storage management system, can realize the logic of storage
virtualization management, more link redundancy management, as well as the
hardware equipment condition monitoring and fault maintenance.
(2) The basic management
Based management is the core part of the cloud storage, is also the most difficult
part of the cloud storage. Based management through cluster and grid computing,
distributed file system such as technology, realize the cloud storage between multiple
storage devices in the work, make multiple storage devices can provide the same
service, and to provide better data access performance, bigger and stronger content
distribution system, 1391, data encryption technology to ensure the data in the cloud
storage will not be access by unauthorized users, at the same time, through a variety
of data for disaster and techniques and measures can ensure that data is not lost in the
cloud storage, ensure the security and stability of the cloud storage itself.
(3) The application of the interface layer
Application of the interface layer is the most flexible part of the cloud storage.
Different cloud storage operation unit can be according to actual business types,
different application service interface, with the application of different services. Such
as video monitoring application platform, network hard disk reference platform, the
remote data backup application platform, etc.
(4) Any an authorized user can access layer through a standard utility application
login interface to cloud storage system, the cloud storage service. Cloud storage
operation services, cloud storage provide different type of access and the access
method.
译文
云存储服务系统研究
Mehra P

摘要
云存储是在云计算 (cloud computing)概念上延伸和发展出来的一个新的概
念，因此,要了解云存储首先 要了解云计算。云计算是一种基于因特网的超级计
算模式，在远程的数据中心里，成千上万台电脑和服务 器连接成一片电脑云。因
此，云计算甚至可以让你体验每秒10万亿次的运算能力，拥有这么强大的计算
能力可以模拟核爆炸、预测气候变化和市场发展趋势。用户通过电脑、笔记本、
手机等方式接入 数据中心，按自己的需求进行运算。随着云计算概念的逐步深入，
人们开始为海量信息寻找一个新的安身 之处一云存储。云存储 (cloud storage)
从一出现就得到广泛的关注和支持。云存储 的概念与云计算类似，它是指通过集
群应用、网格技术或分布式文件系统等功能，将网络中大量各种不同 类型的存储
设备通过应用软件集合起来协同工作，共同对外提供数据存储和业务访问功能的
一个 系统。
关键词:云存储,云存储服务系统,HDFS
1 引言
云存储的兴起使得 整个IT界处于一个重大的变革期，从以设备应用程序为
中心转向以信息为中心，而这一转变还将引发一 系列的变革，并影响到技术和业
务模式两个层面。云存储的最大特点是海量、高性能高访问量以及低成本 ，而
其带来的最大变革是，提供商逐渐从销售工具转向根据工具的实际使用收费，也
就是从卖产 品转向了卖服务。因此，可以说云存储不是存储而是服务。云存储还
具有如下的特点:超强的可扩展性、 不受具体地理位置所限、基于商业组件、按
照使用收费、可跨不同应用等。本文的研究内容为基于HDF S的云存储服务系统
研究，旨在通过构建基于HDFS的云存储服务系统，解决企业的海量数据存储难< br>题，降低实施分布式文件系统的成本，促进Hadoop技术的推广。云存储是在当
前广泛讨论的 云计算概念上延伸和发展出来的，(完整译文请到百度文库）可以



2.4 EMC的策略
EMC推出的云存储基础架构解决方案是一种基于策略的管理系统，让 服务
提供了可以建立不同类别云存储的能力，比如说，其可以为非付费用户创建文件
的两个副本 ，并存储在全球不同的地点，并为付费用户创建5一10份备份进行
存储，并提供了其在全球各地访问文 件的更高的可靠性和更快的访问。在软件系
统中，Atm0S包括数据服务，如复制、数据压缩、重复数 据删除，通过廉价的标
准x86服务器从而获得数百TB的硬盘存储空间。EMC承诺说其拥有自动配置
新的存储空间并自适应硬件故障的能力，也允许用户使用W亡b服务协议进行
管理和读取。目前 Atm0S有三个版本，系统容量分别为 120TB、24OTB和36OTB，
它们全部都基于x8 6服务器并支持千兆或10GB以太网连接。
3 集群存储技术的发展
云存储的兴起正在颠 覆现有的网络存储架构。面对当前PB级的海量存储需
求，传统的SAN或NAS在容量和性能的扩展上 会存在瓶颈。如受到其物理组成
(如:磁盘驱动器的数量、所连接服务器的数量和内存大小和控制器性能 )的限制，
会造成很多功能上的局限(如:支持文件系统的数量，快照或复制的数量等)。一旦
遇到存储系统的瓶颈，就会不断地促使用户升级到更大的存储系统和添加更多的
管理工具，由此增加了成 本。云存储这种新型的服务模式必然要求存储架构保持
极低的成本，而现有的一些高端存储设备显然还不 能满足这种需求。从谷歌公司
的实践来看，它们在现有的云计算环境中并没有采用SAN架构，而是使用 了可
扩展的分布式文件系统GFS)这是一种高效的集群存储技术。GFS是一个可扩展
的分布 式文件系统，用于大型的、分布式的、对大量数据进行访问的应用。它运
行在普通的PC机上，但可以提 供强大的容错功能，可以给大量的用户提供总体
性能较高的服务。云存储130]不是存储，而是服务。 就如同云状的广域网和互联
网一样，云存储对使用者来讲，不是指某一个具体的设备，而是指一个由许许 多
多个存储设备和服务器所构成的集合体。使用者使用云存储，并不是使用某一个
存储设备，而 是使用整个云存储系统带来的一种数据访问服务。所以严格来讲，
云存储不是存储，而是一种服务。云存 储的核心是应用软件与存储设备相结合，
通过应用软件来实现存储设备向存储服务的转变。
4云存储系统分析
与传统的存储设备相比，云存储不仅仅是一个硬件，而是一个网络设备、存
储设备、服务器、应用软件、公用访问接口、接入网、和客户端程序等多个部分
组成的复杂系统 。各部分以存储设备为核心，通过应用软件来对外提供数据存储
和业务访问服务。云存储系统的结构模型 由4层组成。
(1)存储层
存储层是云存储最基础的部分。存储设备可以是光纤通道存储设 备，也可以
是其他的存储设备。云存储中的存储设备往往数量庞大且分布多不同地域，彼此
之间 通过广域网、互联网或者光纤通道网络连接在一起。存储设备之上是一个统
一存储设备管理系统，可以实 现存储设备的逻辑虚拟化管理、多链路冗余管理，
以及硬件设备的状态监控和故障维护。
(2)基础管理层
基础管理层是云存储最核心的部分，也是云存储中最难以实现的部分。基础
管理层通过集群、分布式文件系统和网格计算等技术，实现云存储中多个存储设
备之间的协同工 作，使多个存储设备可以对外提供同一种服务，并提供更大更强
更好的数据访问性能,内容分发系统13 91、数据加密技术保证云存储中的数据不
会被未授权的用户所访问，同时，通过各种数据备和容灾技术 和措施可以保证云
存储中的数据不会丢失，保证云存储自身的安全和稳定。
(3)应用接口层
应用接口层是云存储最灵活多变的部分。不同的云存储运营单位可以根据实
际业务类型，开发不 同的应用服务接口，提供不同的应用服务。比如视频监控应
用平台、网络硬盘引用平台、远程数据备份应 用平台等。
(4)访问层任何一个授权用户都可以通过标准的公用应用接口来登录云存储
系统 ，享受云存储服务。云存储运营服务商不同，云存储提供的访问类型和访问
手段也不同。