Shulou(Shulou.com)06/02 Report--

In the server rental or server hosting host purchase, we will have a server hardware, that is, CPU, memory, hard disk, motherboard and so on, so how do we choose? Let's get to know it with Xiao Yun.

The choice of 1.CPU

As we usually choose to buy PCs, we should first make sure that it is CPU, not the motherboard, because CPU is the core of the entire server. The performance of the service rod depends to a large extent on the type of CPU chosen.

To choose CPU, you first need to know the current server CPU brand, type, and model, or at least the series name. At present, there are only two brands we can choose from in PCs, that is, Intel and AMD, but in terms of server processors, CPU has a much wider range of choices. In addition to Intel and AMO, there are currently IBM and SUN. The processors of Intel and AMO are based on x86 or IA-64 architecture, and the main operating systems supported are Microsoft's Windows system and various brands of Linux systems, while the processors of IBM and SUN are based on RISC architecture and mainly support their own UNIX systems and some brand-specific Linux systems. Generally speaking, Intel and AMO have good processor compatibility and wide support, but generally they can only meet the needs of medium and low-end users. Users such as IDC, telecom, finance, securities and so on usually choose to support UNIX systems, server systems based on IBM and SUN processors. In terms of price, of course, servers based on Intel and AMO processors are cheaper.

Processor branding is only the first step, followed by selecting the type of processor. Among the above four major server processor brands are single processors that support only one processor, dual processors that support only two processors (DP), and of course, there are recently multiprocessors (MP) that support multiple processors. Which one to choose should take full account of the current and future application needs for at least 3 years. If the company is relatively small, and the business development is not very fast, you can consider choosing the instrument to support a single processor, but this is not supported by the author. The author strongly recommends that you choose at least two-way DP processors to maximize the protection of previous investments. However, it is important to know in advance whether the selected DP processor must be installed with two processors (some are possible and some are not) in order to work properly, so as to avoid regret after purchase.

If you choose a processor that supports multiplex, you need to know whether it is a SMTP symmetric architecture or an asymmetric MMP architecture, so that you can purchase the appropriate number of processors according to your actual needs.

After the brand and type of processor are determined, it is necessary to choose a specific model. At present, in the server processors of Intel and AMD, the main competition is dual-core and quad-core processors, coupled with the fierce competition between the two, now the price of dual-core processors is basically the same as that of single-core processors, so when choosing the processors of these two companies, it is best to choose more than dual-core product numbers. For medium-sized and above enterprises, or small enterprises with relatively rapid development, it is recommended to choose quad-core processors, or even multi-core processors that will be put on the market soon. Relatively speaking, processors with multi-core technology are more affordable than buying multiple processors. If quad-core processors are selected and four processors are installed at the same time, the server can allow 16 processes at the same time, which is equivalent to 16 single-core processors in the number of processes.

In addition, there are different series and models of dual-core and quad-core processor products. Which one should be chosen according to the actual demand and economic affordability. It is not necessary to choose the latest one.

IBM and SUN processors are currently typically dual, quad-core and eight-core processors, mainly used to support their own UNIX operating systems.

two。 The choice of motherboard

After selecting the processor, the next thing to do is the choice of the server motherboard, which is the most direct embodiment of the new technology, new functions, new features and comprehensive performance of the whole server. The motherboard of the server is more important than the motherboard of the ordinary PC machine, because the main meal in the server not only determines the latest technology adopted by the server, but also determines the overall performance of the service. Of course, the choice of the server motherboard is first of all decided by the processor choice that is proud of the front of the chimney. But the server is not like PC, we choose the motherboard in the market, but buy it in the form of the whole machine. At this time, the performance of the server motherboard is obtained through the introduction of the corresponding model of the server. It is usually recommended to choose the raw wrench chip developed by the corresponding brand processor manufacturers. The thin motherboard makes it more compatible and can maximize the new technologies and features of the corresponding processors.

In choosing the motherboard, we mainly consider the standard configuration and extended performance of the server. Such as the number of processors, the type and maximum amount of memory supported, various disk interfaces, various expansion slot types and numbers (such as PCI, PCl-X, PCI-E or lnfiniBand Imax O, etc.), whether to integrate RAID control cards, RAID control card types, integrated ask card type and quantity, and some other internal / external interface configurations.

The above configuration depends on the server level selected, the entry-level server configuration is the lowest, the processor generally supports only 2 channels, the memory usually only supports 4GB, and the disk rack usually provides only 4 slots, the number of Igamot O expansion slots is usually less than 6 (usually using 32-bit and 64-bit PCI bus), does not support external expansion. Suitable for small businesses with about 50 users.

The configuration of the departmental server is slightly higher than that of the workgroup server, the processor supports 4x8, the maximum memory 12GB is about 8, the number of disk racks is about 8, and the number of expansion slots is about 8 (usually provides new buses such as PCI-X and PCI-E), and usually supports external extension. It is suitable for medium-sized enterprises with more than 200 users and rapid business development.

Enterprise-class servers have the highest configuration, usually supporting more than 8 processors, and the maximum memory is about 16GB. More than 10 disk shelves can be provided through external cabinet expansion, and a variety of Imax 0 expansion slots (usually provide new buses such as PCI-X, PCI-E, lnfiniBand) can also be extended through peripheral expansion. Suitable for enterprise users with more than 500 users.

3. The choice of memory

Memory is second only to CPU, especially in servers, because many server failures are caused by memory. In terms of server memory selection, we should not only consider the current mainstream memory technologies, such as DDR2, DDR3, and the amount of memory it supports, but also consider its other features, such as supported expert modules (such as ordinary DIMM, and now the latest FB-DIMM), error correction methods (at least to support standard ECC), whether to support internal mirroring, memory protection, memory array and other technologies.

4. Selection of hard disk

Because the server has to face multi-user concurrent file access, the hard disk used by the server is required to have very high read and write performance. On the one hand, it is guaranteed by improving the speed of the hard disk and the bandwidth of the hard disk interface, on the other hand, it should be strengthened by magnetic removal array (RAID).

At present, in terms of hard disk interface, the server hard disk mainly uses parallel SCSI, but the new SATA and SAS interfaces are also beginning to be used in some high-end servers, because these two new interface technologies have become very mature after years of development. The speed of the hard disk is usually above IOOOOPPM.

In terms of RAID, SC buttons, RAID, SATA-RAID and SAS-RAID can also be used in the server field, which is consistent with the hard disk interfaces of the three main applications. In contrast, SCSI-RAID and SAS-RAID have higher performance, and the interface technology of SAS is relatively newer. At present, there is less server support and less choice of this kind of array cards.

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.