1, the server is designed according to the 24-hour long-term operation mechanism, which is more stable and of course more expensive.
2. As for the data bus, the server is faster and generally adopts SCIS or SAS.
3. The server has RAID function and supports disk array.
If you don't have high requirements for the machine that makes the server, a general PC can do it; Conversely, if you use the server as a PC, you will feel out of place.
Specifically, you should know the difference between a server and an ordinary PC, which depends on the hardware and software.
According to different applications, there are great differences between them. For example, a PC is equivalent to an outpatient doctor in any city, but its medical skills are not so superb. The server should be an expert in a certain field. The more outstanding its processing power, the more powerful it is.
First, from the hardware point of view:
1. CPU
The instruction of the server CPU is generally RISC (Reduced Instruction Set). According to the research, in most applications, CPU only uses few commands, so researchers designed instruction sets according to this situation, and used a variety of centralized command combinations to meet various requirements. The advantage of this design is that it is more targeted, can be specially optimized according to different needs, and has higher processing efficiency. Corresponding to CISC (Complex Instruction Set), it is characterized by integrating various common functions as much as possible, such as MMX, SSE, SSE+, 3D! Now! Wait, it's all this type.
In addition, the CPU design of the server should generally consider its multi-channel function, that is, multiple CPUs work together, while the PC generally has only one CPU.
Step 2 remember
In principle, the larger the memory on the server, the better, but the requirements for error correction and stability are higher, such as ECC ("error detection and correction" seems to be nobody called). Few people can use 1G memory on the PC we are using now (not counting playing games), but on the server, this G-level memory is sometimes obviously stretched. I remember that when the country released the latest supercomputer of Galaxy last year, the memory reached1t; Compared with the speed of memory, people give priority to the stability and error correction ability of memory in application. Only by ensuring these two points can we consider the others.
3. Hard disk
The performance of hard disk, whether on PC or server, has been slowly improving. Personally, relying on the development of machinery, the development of hard disk can not have a qualitative leap. Because servers are generally used by enterprises, which store a lot of precious data, this puts forward the requirements of security and stability for hard disks, and the related technologies appearing on hard disks are basically around these two requirements. For example: redundant data backup, hot plug, etc. In addition, the server hard disk must be able to work continuously 24*7.
4. Motherboard
I don't know much about this, and I rarely see the saying that the server has a motherboard, but I think it's necessary to mention the bus design of the server-multi-channel, that is, how multiple CPUs work together. If you are interested, I suggest you read books on operating systems, written by foreigners, which are very good!
5. Graphics card
In addition to graphics and 3D design, the graphics card on the server can basically be displayed as long as you can connect the monitor!
Second, talk about software. Software mainly refers to operating systems, such as NT, 2000 SERVER, 2003 SERVER, LINUX, SOLRAIS and UNIX, which are specially designed for servers, such as load balancing and multi-channel CPU support.