You only need to take a high-quality server. Because he is a keeper, he is also a liaison. He – as in that proverb: “and a priest, and a reaper, and to his happiness a blacksmith.” Read more about how to choose the right server in this article.
For what purposes is the server chosen?
Servers are used for any activity – from storing client data to connecting players during a game battle.
It all depends on the configuration: the more load on the server, the more powerful it should be. Usually, it is purchased for one of the following purposes.
- Mail Server
Used to work with e-mail. The device processes the letter before it reaches the addressee.
2. File
It is like Plyushkin – it stores files on the network to which other computers are connected.
3. Web server – buzzer in the server world
It is constantly online, accepting user requests.
4. Game server is a must for gamers
Users connected to it can interact with each other on the game map or online game site. They need huge computing power, since even fractions of a second are important here.
5. A proxy server is a sister, but not a twin of a web server
This is a set of programs that makes requests on the web instead of the visitor. The user gives a request to the proxy server, which looks for the necessary information on its own.
Almost every company has a database server. Many work programs use databases to run, which can take up terabytes of hard disk space. And in order not to overload computers with a constant connection with each other, the databases are loaded onto the server. And they all take information from there, without rummaging in a neighbor’s computer.
Educational program: How to choose a compatible motherboard and video card: a guide in 4 sections
Floor Server
The tower case is suitable for those looking for a compact budget server that can be located in the office. The simplest floor – standing models are suitable even for a home server. The performance is adequate – enough for a few people.
These devices are good for small business databases. If the needs increase, you will have to buy more “towers”, because of which they will take up a lot of space. They are not suitable for a rapidly developing business – there may not be enough space in the office.
| pros | Minuses |
| Easy to replace the parts you want | Scales poorly |
| Quiet | – |
| Unpretentious | – |
Rack server
The most powerful server type. For this, you will have to allocate a special server room, as it will be noisy due to the enhanced cooling system. New elements are added to the rack as easily as puzzles in a construction set.
The price of such models “bites”. The cost of only one server rack is the same as the “floor” model. However, if the company is constantly increasing capacity and often buys parts, he will quickly beat himself off.
A rack server is the choice for large companies that have to solve thousands of costly tasks at the same time.
| Dignity | Shortcomings |
| Easy to expand | Noisy |
| Reliable and functional | – |
On the intricacies of work: Two video cards in a PC: good or bad – 3 answers
Blade server
Blade is translated from English as “blade”. Such devices are compact and easy to install. The key difference between “rack” and “blade” is that the latter are installed in the chassis / chassis. Adding a new part will be as easy as inserting the blade into the chassis. So here you can create any configuration, while not worrying about the wires. However, due to the additional infrastructure, the initial cost of the kit will be higher.
At the same time, they have fewer opportunities for expansion, since the “blades” are not equipped with such a volume of PCI Express slots and cages for drives, as other servers.
| Dignity | disadvantages |
| High performance | Fascinating price |
| Ease of installation | Fewer options for expansion – you will have to choose consumables carefully |
| Scalability – you can easily expand server capacity without cluttering up the space around | – |
Read: How to Start a Motherboard Without a Button – 3 Ways
Particular attention to components
When buying or compiling a server, pay attention to the components. It depends on them how much the device can “pull” and what loads it is guaranteed to withstand.
CPU
Typically, the processor is matched to the size of the workload and the number of requests that will have to be handled. This is an expensive technique, it is immediately taken “for growth”: so that it will suffice, taking into account the future increase.
Server percent is characterized by two main parameters: frequency and number of cores. Both should be on top. The first characteristic is important when the system loads one core.
But in multithreaded processes, it is important how many cores the system can load. Then they will not work “at full capacity”, which means they will live much longer.
As of mid-2020, the average number of cores in servers for small offices should be 6-8. A large office needs at least 10 cores, or even more.
We also recommend that you pay attention to three not critical, but still important parameters.
- Cache size
Affects the speed of work. The bigger, the better.
- Socket type
The connector through which the percent is connected to the motherboard. The use of such connectors instead of soldering the core to the “motherboard” simplifies the replacement of the processor in future upgrades and significantly reduces the price of the motherboard. Choose a processor with the same socket type as your motherboard.
- Clock frequency
The greater the load, the higher this parameter should be. For a server to which several computers are connected, 2-2.5 GHz is enough. But the equipment that controls a dozen PCs, cameras and other “smart” devices, you need to look for processors with a CPU frequency of 4-5 GHz.
On the topic: CPU and motherboard compatibility – how to choose components: a guide in 3 sections
RAM
The more RAM, the better, but this way server modules will cost more than standard ones for a PC. It is better to calculate the required volume. If averaged and on a scale of numbers, for every 100 people of staff, about 8 GB of RAM will be needed. Its number may vary depending on the complexity and number of tasks.
For an accurate calculation of RAM, it is easier to use the formula:
X = 256MB + 64MB * Y + 0.5 * Z
- X – the required volume of the operative;
- Y is the number of staff;
- Z is the volume of the database with which the personnel works.
Real life example: you need to find RAM for a company of 75 people working with a 15 GB database. As a result, it will be 12.5 GB.
Everything was calculated like this: 256 MB + 64 MB * 75 + 0.5 * 15360 MB = 256 MB + 4800 MB + 7680 MB = 12736 MB ≈ 12.5 GB.
Since the server technology is taken “for growth”, it is better to equip the server with 14-15 GB. There are modules from 256 MB on the market, so it will not be difficult to do this.
You should also consider the memory standard. As of mid-2020, there are 2 main modulations on the market – DDR3 and DDR4. The key difference between them is in the speed of information transfer. In the first case, it is 1333-1867 MHz, and in the second 2133-2666 GHz. The higher the number, the more efficient the server will be.
Choosing a server “for the future”, it is better to install DDR4 right away. It is more modern, while the “troika” is becoming a thing of the past.
Interesting article: How to increase the random access memory (RAM, RAM) of a computer in 4 steps – all the ways
Disk subsystem
The disk subsystem directly affects server performance. Even if the unit has several free gigabytes, but the hard disk is selected incorrectly , the server will be noticeably dull.
As of mid-2020, the servers are equipped with disks with the interfaces listed in the table:
| Disc type | Characteristic |
| SATA | With a large volume, it lives for a long time, but the speed of access to information is low |
| SCSI | Differs in durability and high speed of work with data |
| Fiber Channel | Ideal for web and proxy servers that need to respond to multiple requests at once |
| SAS | Disk system, which allows you to equip the server with two types of “hard”. For example, you can store basic information on SATA, and write frequently used data into fast SAS. |
Hard type can be either HDD or SSD. The former will withstand more data rewriting processes, the latter will be noticeably faster.
Arrangement for correct processor operation
Once you’ve found the basic details, it’s time to start thinking about how to properly set up the server itself. It will work all the time, so you need to take care of the distribution of power and its continuity. A lot of important information is stored on the server, so sharp metamorphoses will only harm the device.
Plus, dozens, or even hundreds or even thousands of people constantly connect to the servers. Consequently, he has to constantly work, and without a well-thought-out cooling system, he will overheat and shut down forever. Below in the text – how to correctly configure the processor at the proper level.
Educational article: What are timings in RAM, which are better – educational program in 4 sections
Power distributor
As the name implies, this is a special device for distributing electricity. With it, each part of the server receives exactly as much energy as it needs.
The owner of the distributor can remotely control the device down to individual sockets. In addition, it has five useful functions:
- limiting voltage;
- remote control of sensors connected to the network;
- control of sockets and the ability to increase their number thanks to special modules;
- accurate measurement and fixation of power and current in the network;
- the presence of connectors that allow you to connect anything: from sensors to webcams.
The distributors themselves are divided into simple (without controls) and controlled (which is administered remotely).
On the topic: What connectors are on the motherboard and what are their names: educational program in 4 sections
Uninterrupted power supply unit
A system that will provide the connected equipment with current in case of a short-term disconnection of the main power source. Also, the UPS protects against disturbances in the network and stabilizes the supplied electricity.
Such a system protects the server from current surges and will allow the connected devices to be maintained for a long time in the event of a network breakdown. The source type is selected depending on the size and capacity of the server. There are 4 types of them:
- Static (SIBP)
Built on a rechargeable battery that is charged from the mains + inverter.
2. Diesel generator (DGI)
Starts when the mains is lost.
3. Dynamic (DIBP)
This is a motor-generator with a mechanical battery. Its advantage is that there will be no interference, since it outputs a pure sine wave to the load.
4. Diesel-dynamic (DDIBP)
A reliable and simple choice that combines the advantages of all UPSs.
Cognitive article: What kind of battery is on the motherboard and how to replace it – 4 important nuances
Cooling system
Constant use puts stress on computer components, causing them to overheat and fail. Therefore, when designing a server, you need to think about a cooling system.
There are three types of cooling used in computer technology.
- Air – fans that, at peak loads, make noise like party-goers at a rave.
- Liquid (dropsy) – a working fluid that flows through special pipes, receives the heat generated by the components. Then it enters the condenser, where it is cooled and re-launched into the system. Distilled water or special safe mixtures are used as a liquid.
- Combined – such servers use several cooling systems at once.
Before buying a device, you need to study the characteristics in detail or trust the professionals. Otherwise, the process of replacing parts will further increase the cost of the final price of the finished server.
