普通集群 镜像集群_集群| 第2部分

普通集群 镜像集群

集群发展 (Development of Clusters)

The initial business oriented solution of the cluster has been ARCnet, designed by Datapoint in the year 1977, however It was not until 1984 when VAXcluster produce the operating system VAX / VMS. The ARCnet as well as VAXcluster are not just the kind of solutions which help parallel processing, but additionally help tremendously in shared file systems as well as peripheral devices.

群集的最初面向业务的解决方案是由Datapoint在1977年设计的ARCnet，但是直到1984年VAXcluster才生产操作系统VAX / VMS。 ARCnet和VAXcluster不仅是有助于并行处理的解决方案，而且在共享文件系统以及外围设备中也提供了极大的帮助。

The concept seemed to be to deliver the conveniences of parallel processing, while still be able to maintain the crucial data reliability as well as uniqueness. VAXcluster, VMScluster continues to be offered with HP OpenVMS systems operating on Itanium as well as Alpha systems.

这个概念似乎是为了提供并行处理的便利，同时仍然能够保持关键的数据可靠性和唯一性。 VAXcluster，VMScluster继续与在Itanium上运行的HP OpenVMS系统以及Alpha系统一起提供。

Two additional significant features of the original commercial clusters have been the Tandem Himalaya and also the IBM S/390 Parallel Sysplex that was used solely for business purpose.

原始商业集群的另外两个重要功能是Tandem Himalaya和仅用于商业目的的IBM S / 390 Parallel Sysplex。

The historical past of computer clusters can be incomplete without writing about the critical role performed by the software development of the Parallel Virtual Machine (PVM). This particular open source software dependant on TCP / IP helped in the development of a virtual supercomputer – an HPC cluster – built from any of the systems that were connected through the TCP / IP.

如果不编写并行虚拟机(PVM)的软件开发所扮演的关键角色，则计算机群集的历史记录可能是不完整的。 这种依赖于TCP / IP的特定开源软件帮助开发了虚拟超级计算机(HPC群集)，该超级计算机由通过TCP / IP连接的任何系统构建而成。

PVM and also the utilization of Personal computer as well as inexpensive networks contributed in 1993 to a NASA venture to construct supercomputers by using clusters. These advancements additionally stimulated the independent progression to the Grid computing, in spite of the Grid style revolved about the Unix OS (operating system) and also the Arpanet.

PVM以及个人计算机以及廉价网络的使用在1993年为NASA冒险通过使用集群构建超级计算机做出了贡献。 尽管Grid风格围绕Unix OS(操作系统)和Arpanet，但这些进步还促进了Grid计算的独立发展。

集群技术的优势 (Advantages of the Cluster Technology)

Scalable parallel software programs need: great overall performance, low latency, highly accessible communications bandwidth, scalable network systems, and also a very fast access to information. A cluster satisfies all these specifications by making use of all the great resources it enjoys.

可扩展的并行软件程序需要：出色的整体性能，低延迟，高度可访问的通信带宽，可扩展的网络系统以及对信息的快速访问。 集群通过利用其所拥有的所有巨大资源来满足所有这些规范。

Clusters provide these functions at a fairly affordable rate:

集群以相当实惠的价格提供以下功能：

• High performance
  高性能
• High availability
  高可用性
• High efficiency
  高效率
• Scalability
  可扩展性

Cluster systems enable businesses to improve their own processing capability by making use of standard technologies, both computer hardware as well as computer software that could be obtained at a reasonably low cost.

集群系统使企业可以利用标准技术(包括计算机硬件以及可以以较低成本获得的计算机软件)提高自身的处理能力。

集群分类 (Classification of Clusters)

The word cluster offers diverse connotations for various groups of people today. The kinds of clusters, set up according to the use and also the services they provide, define the actual significance of the term for that group which utilizes it. The clusters may be categorized based on their features: they could possess high-performance clusters (called HPCC – High Performance Computing Clusters), clusters associated with high availability (HA – High Availability) or clusters of high efficiency (HT – High Throughput).

词簇为当今的不同人群提供了多种含义。 根据用途及其提供的服务建立的群集类型，定义了该术语对使用该术语的组的实际意义。 这些群集可以根据其功能进行分类：它们可以具有高性能群集(称为HPCC –高性能计算群集)，与高可用性关联的群集(HA –高可用性)或高效群集(HT –高吞吐量)。

High performance: They are clusters where activities are performed which need high computational capability, substantial memory, or even the two simultaneously. Performing these types of work may possibly endanger the cluster resources for extended amounts of time.

高性能：它们是执行活动的群集，这些活动需要较高的计算能力，大量内存，甚至同时需要两者。 执行这些类型的工作可能会长时间危害群集资源。

High availability clusters were created in order to offer availability as well as reliability. These types of clusters making the effort to offer maximum availability involving their services. The actual reliability can be offered by software program which finds problems and permits them to recover from, although the computer hardware is actually avoided.

创建高可用性集群是为了提供可用性和可靠性。 这些类型的集群努力提供涉及其服务的最大可用性。 实际的可靠性可以由软件程序提供，该软件程序可以发现问题并允许他们从中恢复，尽管实际上可以避免使用计算机硬件。

High Efficiency: These types of clusters whose creation was to enable running of as many functions within the least time period.

高效率：这些类型的集群的创建是为了在最短的时间内运行尽可能多的功能。

集群的组成部分 (Components of a Cluster)

In general, a cluster needs several software and hardware components to function. Namely:

通常，集群需要几个软件和硬件组件才能运行。 即：

• Node
  节点
• Storage
  存储
• Operating Systems
  操作系统
• Network Connections
  网络连接
• Middleware
  中间件
• Communication protocols and services
  通讯协议和服务
• Applications
  应用领域
• Parallel Programming Environments
  并行编程环境

Nodes

节点数

Computers can be simple, multi-processor systems or workstations (workstations). In computing, a very general way, a node is a point of intersection or union of several elements that come together in the same place. However, within the word computer node can refer to different concepts depending on the area in which we move:

计算机可以是简单的多处理器系统或工作站(工作站)。 在计算中，一种非常通用的方式是，节点是在同一位置集合在一起的多个元素的交点或并集点。 但是，根据我们所移动的区域，“计算机节点”一词可以指代不同的概念：

• In computer networks each machine is a node, and the network is the Internet, each server is also a node.
  在计算机网络中，每个机器都是一个节点，而网络是Internet，每个服务器也是一个节点。
• In dynamic data structures a node is a record that contains a point of interest and at least one pointer to reference (point) to another node. If the structure has only a pointer, the only structure that can be built with it is a list, if the node has more than a pointer and can build more complex structures such as trees or graphs.
  在动态数据结构中，节点是一条记录，其中包含一个关注点和至少一个指向另一个节点的参考(指向)指针。 如果该结构只有一个指针，则该节点可以建立的唯一结构是列表，如果节点具有多个指针并且可以建立更复杂的结构(例如树或图)。

The cluster nodes may comprise nodes involved or not involved.

群集节点可以包括涉及或不涉及的节点。

In a cluster with dedicated nodes, the nodes do not have a keyboard, mouse or monitor and its use is exclusively dedicated to performing tasks related to the cluster. While in a non-dedicated cluster with nodes, the nodes have a keyboard, mouse and monitor and its use is not solely dedicated to tasks related to the cluster, the cluster makes use of clock cycles that the user’s computer does not being used to carry out their tasks.

在具有专用节点的群集中，这些节点没有键盘，鼠标或监视器，其使用专门用于执行与群集相关的任务。 在具有节点的非专用群集中，节点具有键盘，鼠标和监视器，其使用不仅专门用于与群集相关的任务，该群集还利用了用户计算机不用来承载的时钟周期完成任务。

It is clear that when designing a cluster, nodes must have similar characteristics, i.e., they must bear some similarity of architecture and operating systems, because if it forms a fully heterogeneous cluster with nodes (there is a big difference between capacity processors, memory, hard drive) will be inefficient because the middleware delegate or assign all the node processes more computing power and distributed only when this process is saturated, so it is recommended to build a group of computers as similar as possible.

显然，在设计集群时，节点必须具有相似的特性，即它们必须具有某种架构和操作系统的相似性，因为如果它与节点形成完全异构的集群(容量处理器，内存，硬盘驱动器)效率低下，因为中间件委派或分配所有节点进程更多的计算能力，并且仅在此进程饱和时才分配它，因此建议构建一组尽可能相似的计算机。

Storage

存储

The storage may be a NAS, SAN, or internal server storage. The protocol most commonly used is NFS (Network File System), file system shared between server and nodes. However, there are specific file systems like Lustre clusters (CFS) and PVFS2.

该存储可以是NAS，SAN或内部服务器存储。 最常用的协议是NFS(网络文件系统)，即服务器和节点之间共享的文件系统。 但是，有特定的文件系统，例如Lustre群集(CFS)和PVFS2。

Technologies in support of hard disk storage:

支持硬盘存储的技术：

• IDE / ATA speeds of 33, 66, 100, 133 and 166 MB / s
  IDE / ATA的速度分别为33、66、100、133和166 MB / s
• SATA speeds of 150, 300 and 600 MB / s
  SATA速度为150、300和600 MB / s
• SCSI speeds of 160, 320, 640 MB / s. Provides high yields.
  SCSI速度为160、320、640 MB / s 提供高产量。
• SAS: combines SATA-II and SCSI. Speeds of 300 and 600 MB / s
  SAS：结合了SATA-II和SCSI。 300和600 MB / s的速度
• Tape drives (DLT) are used to back up for their low cost.
  磁带驱动器(DLT)用于备份它们的低成本。

NAS (Network Attached Storage) is specifically dedicated to the storage device through the network (usually TCP / IP) that uses an operating system optimized to provide access via CIFS, NFS, FTP or TFTP.

NAS(网络附加存储)专用于通过网络(通常是TCP / IP)的存储设备，该网络使用经过优化的操作系统来提供通过CIFS，NFS，FTP或TFTP的访问。

For its part, DAS (Direct Attached Storage) is to connect external drives SCSI storage or a SAN (Storage Area Network) via a Fibre Channel. These connections are active.

就DAS(直接连接存储)而言，它是通过光纤通道连接外部驱动器SCSI存储或SAN(存储区域网络)。 这些连接处于活动状态。

While NAS to share storage, using the network, and has a simplified management, DAS provides improved performance and reliability by not sharing the resource.

NAS通过网络共享存储，使用网络并简化管理时，DAS通过不共享资源来提高性能和可靠性。

Operating Systems

操作系统

An operating system must be multi-threaded and multi-user skills. Other desirable features are ease of use and access. An operating system is a program or set of computer programs designed to ensure effective management of its resources. An operating system starts working when some one turns on the computer, and it then manages the hardware of the machine from the most basic levels, allowing the interaction with the user. Operating system can usually be found in most electronic devices that use microprocessors to operate, and that through these we can understand the machine and that it fulfills its functions (cell phones, DVD players, radios, computers, etc.). Linux, Windows, UNIX, Mac OS X are some of the very popular operating systems that have revolutionized the online world.

操作系统必须具有多线程和多用户技能。 其他理想的功能是易于使用和访问。 操作系统是旨在确保对其资源进行有效管理的一个程序或一组计算机程序。 当有人打开计算机时，操作系统便开始工作，然后从最基本的级别管理计算机的硬件，从而允许与用户进行交互。 操作系统通常可以在使用微处理器进行操作的大多数电子设备中找到，通过这些操作系统，我们可以了解机器并实现其功能(手机，DVD播放器，收音机，计算机等)。 Linux，Windows，UNIX，Mac OS X是彻底改变在线世界的一些非常流行的操作系统。

Continued…

继续…

翻译自: https://www.eukhost.com/blog/webhosting/clusters-part-2/

普通集群 镜像集群