Shulou(Shulou.com)05/31 Report--

This article is about how to select storage devices in the IT project. I think it is very practical, so I share it with you. I hope you can get something after reading this article. Let's take a look at it.

In the IT project, because the storage equipment involves a relatively large scale of investment and a long life cycle, it is more important. So how to select the storage device in the IT project? Next, we will make an analysis and explanation from the following aspects.

There are several factors that must be taken into account when choosing the right storage device: protocol, capacity, performance, scalability, manageability, and cost. In addition to the determinants of these standards, you should also consider that the selected storage device will not be in a vacuum outside the network, computing, and monitoring management platform. Storage devices must be seamlessly integrated with the IT infrastructure as much as possible.

I. Agreement

Protocol is the first problem to be considered in the selection of storage devices. Choosing which storage protocol plays an important role in the overall decision of the IT project, because it can and will continue to decide which vendor's products and platforms to use. In addition, it affects the overall design, architecture, and type of infrastructure. Fibre Channel storage network is currently the most widely used storage network in enterprise data centers. FC protocol and iSCSI protocol play an important role. In fact, FC protocol can not be translated into optical fiber protocol, but FC protocol generally uses optical fiber as transmission cable rather than copper cable, so many people call FC fibre Channel protocol. Logically, we can think of FC as a bi-directional, point-to-point serial data channel for constructing high-performance information transmission. Physically, an FC is a point-to-point interconnection link corresponding to one to more, with each link terminating at a port or transponder. The link media for FC can be optical fiber, twisted pair, or coaxial cable.

ISCSI technology is a kind of SCSI instruction set researched and developed by IBM Company, which can be used by hardware devices and run on the upper layer of IP protocol. This set of instructions can run SCSI protocol on IP network and enable it to choose routes on high-speed Gigabit Ethernet. ISCSI technology is a new storage technology, which combines the existing SCSI interface with Ethernet technology, so that the server can exchange data with the storage device using IP network. Moreover, ISCSI protocol reduces the deployment cost of the system and solves the problems of compatibility and unified management, which represents the future of storage development.

In addition, the fibre Channel over Ethernet (FCoE) standard has been in shape, the product has been on the market for several years, and the end-to-end system is mature. The FCoE technology standard maps fibre Channel to Ethernet and inserts fibre Channel information into Ethernet packets, allowing fibre Channel requests and data from server-SAN storage devices to be transmitted over Ethernet connections without the need for a dedicated fibre Channel structure, allowing SAN data to be transmitted over Ethernet. FCoE allows LAN and FC SAN communications to be transmitted on a single communication cable, and converged networks can support LAN and SAN data types, reducing the number of data center equipment and cables, while reducing power and cooling loads, converging into a unified network, reducing the number of support points and helping to reduce the administrative burden. As an industry, because of the many advantages of network consolidation, it is mainly developing in the direction of Ethernet or IP protocols (iSCSI and FCoE), but fibre Channel still has a large market share because it still has great advantages in traditional investment, performance reliability and customer trust. Generally speaking, the storage architecture of the whole project needs to be considered by combining the network storage structure (DAS, NAS, SAN).

Comparison of storage protocols

II. Capacity and performance

Capacity and performance are also important considerations for customers. The maximum disk capacity of traditional storage systems is determined by the number of disks they support and the size of those disks. In other words, the maximum capacity usually confuses customers, because without features such as tiering and caching, performance will be linked to the number of disks. So you have to have more disks to ensure IOPS performance, and a lot of spare storage capacity that is not being used.

You want to choose a storage platform that can be flexibly expanded to meet your capacity needs during the infrastructure cycle. A better method is to select a cluster composed of several "general storage devices" for storage, and the performance and capacity of each storage system that makes up the cluster storage can be superimposed and expanded by the way of "cluster". Of course, even if you use a storage cluster, once you encounter a bottleneck in the storage system, there are two options: one is to use a single storage system with more powerful hardware; the other is to use several common performance storage systems to form a "storage cluster". This provides a proportional increase in the performance, capacity, reliability and availability of storage resources, breaking through the limitations of stand-alone devices.

Performance is significant for any storage platform. In a typical private cloud deployment platform, performance is divided into multiple levels, which can generally be divided into silver, gold and platinum levels. You want your storage platform to have multiple service levels without running out of resources (that is, you don't want to spend on top-level services in each service). Some functions of storage, such as automatic tiering (arrays frequently read data from disk, or vice versa), and computer cache processing (front-end cache strips of memory), can achieve high performance at low consumption. In addition, features such as data deduplication, replication, and snapshots are also worth considering.

III. Scalability, manageability and cost

Expandability

In general, cluster storage should include three elements: storage node, front-end network and back-end network. each element can easily adopt the latest technology in the industry without changing the architecture of cluster storage, and it is very convenient to expand. expand storage like building blocks. Especially for those users who are difficult to predict the data growth trend, they can buy part of the storage first and add it at any time when there is a demand, without affecting the use of existing storage.

Distributed storage can improve the performance of the system, especially the scalability. All the operations on cluster storage are uniformly scheduled and distributed by the distributed operating system, and are distributed to each storage node of cluster storage. The advantage of using distributed operating system is that there is no difference between each node, no primary or secondary, no functional difference, all storage nodes have the same function, so that the performance can be the best.

Manageability

At present, the management methods of the storage industry are through the management tools of various manufacturers, or through the Web interface for management and configuration, often the client also needs to install relevant software to access the storage space. As the storage space to be managed increases, so does the complexity of managing storage and the number of managers. Cluster storage should provide a centralized, easy-to-use management mode without any impact on clients, and use industry standard access protocols (such as NFS,CIFS) to access cluster storage.

Cost

Cost is a key factor in any IT project. For a private cloud, you must need a system that can be managed and detected by both local tools and your own automated monitoring platform. The more open the storage management platform is, the more powerful its application interface or software system will be, and the more convenient it will be to use. When you make a decision, you must make sure that the software you choose for the automated monitoring and management platform can well adapt and manage your storage platform. These tools are often used and expensive, and it is important that they limit your choice of hardware.

For an IT project, the choice of storage devices is complex and important. Even if you don't think about virtualization now, you have to think twice about your choices in order to develop your storage platform in the cloud in the future. And it is a serious challenge for all enterprises to manage storage and data well with the increasing complexity and risk of IT system and the rising cost of procurement and management.

In a nutshell, the storage needs of enterprises are more or less the same, which can be briefly summarized as follows:

1. Effective control of total cost of ownership (TCO)

2. Security, reliability and consistency of data

3. Provide outstanding high-end storage services at low cost

4. Realize the integration of different storage levels.

5. Reduce the complexity of storage management

6. Storage scalability and high performance

7. Data storage life cycle management.

The above is how to select the storage devices in the IT project, and the editor believes that there are some knowledge points that we may see or use in our daily work. I hope you can learn more from this article. For more details, please follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.