Unfortunately, not all managers know the difference between a regular work computer and a server. It is often quite difficult for a system administrator to explain to his boss why, for the normal operation of the corporate network, it is impossible to pull the old system unit out of the closet, but it is necessary to purchase a server dell, HP or any other well-known manufacturer.
Functionality
Despite the external similarity, the PC and the server have completely different tasks. The latter is necessary, first of all, in order to distribute the resources of office computers in the local network. Yes, there is no difference in hardware. And if we are talking about a small number of machines, then these functions can be assigned to a standard PC. But even in this case, it is necessary to install a special server version of the operating system and a number of special programs on it.
But as soon as the network grows, the load on the computer increases significantly. And in this case, whether you like it or not, you will have to buy a server whose price, by the way, is not so high.
Fundamental differences
First of all, it is necessary to note the more powerful "stuffing", mainly motherboards. These computers must perform incomparably more processes than office machines. Again, due to the heavy load, the servers are equipped with multiple power supplies.
In their work, a high speed of connection to the network is extremely important. The functioning of the entire system depends on this. In addition, special software is demanding in terms of resources. That is why the standard amount of RAM for a server today is about 64 GB. And the capacity of hard drives is in the hundreds of terabytes.
In a word, a server is a very powerful computer on which special software is installed. By and large, it performs only one function, but if it “falls down”, the entire network connected to it will stop working.
There are very few publications about servers and server hardware. And the main reason is the technical complexity - there are many differences from the usual consumer hardware, and a limited readership. Such articles are of interest only to administrators and those who make purchasing decisions, and to some enthusiastic readers who are fond of professional-grade hardware. However, server hardware is closer to desktop hardware than you might think, and more knowledge never hurts.
When people think of servers, they think of big computers, heavy boards, and outrageous performance, but the reality is often different. Today there are so many form factors and a huge amount of hardware and software that it is difficult to come up with a universal definition of the word "server".
While there are many similarities between professional and consumer hardware, we believe that it is the emphasis on certain features and qualities that makes it possible to classify the hardware as professional. For example, your home PC needs to be fast, quiet, upgradable, and, of course, reasonably priced. It will work for several years, while it will often be idle for several hours, and the user will have the opportunity to replace the failed hardware or simply remove the accumulated dust. Other requirements are imposed on servers: reliability, 24/7 availability, maintenance without stopping work are in the first place here.
First and foremost, the server must be reliable. Whether it's a database server, a file server, a web server, or another type of server, it must be very reliable because your business depends on it. Secondly, the server must always be available, that is, the hardware and software must be selected in such a way that downtime is minimal. Finally, fast maintenance in the professional field is very critical. That is, if an administrator needs to perform a task, it must be performed as efficiently as possible without conflicting with the criteria mentioned above. That is why server performance is often the result of consideration of the necessary requirements and long-term strategies, and not the result of some emotional step, as is often the case with gaming PCs.
In this article, we will talk about server components and describe the technologies that are common to servers and consumer PCs, as well as talk about the differences and advantages. Since all professional-grade components are much more expensive than ordinary ones, we will begin our digression with this question.
Professional means expensive
If you buy professional components or servers and workstations, you will quickly find that they cost more than regular consumer hardware. And the reason often lies not in some complex technology, but in the specifications of professional components, in their testing and validation. For example, the Core 2 Duo Conroe is very close to the Xeon Woodcrest in terms of performance. But the differences lie in the sockets used, specifications, and systems in which these processors are installed. Server hard drives are specifically designed for continuous 24/7 operation, while desktop hard drives are not.
Usually we assume that any consumer products are compatible with all others, which is not always the case, but most often. Therefore, you can replace one compatible component with another, most likely there will be no problems. But this approach is no longer acceptable if you are planning to upgrade the server or perform maintenance.
New products for the professional market are developed with a predictable upgrade path as manufacturers want these products to work with existing systems, current and future generations of components. AMD and Intel customers regularly receive company plans for their products that provide a glimpse into the future. Consumers can buy a product with the confidence that they will receive support and upgrade opportunities for a while.
Warranty and replacement of components is also very important. If a failed desktop hard drive is replaced under warranty with any new model, then professional solutions often require exactly the same components. Therefore, the administrator needs to look for exactly the same product, while ordinary users, on the contrary, will be unhappy if they do not get the latest generation of components (which, by the way, is cheaper for most manufacturers).
The magic word for the professional market is validation. When a groundbreaking new product is ready for release, it will be validated and tested on popular hardware systems. The validation process ensures that companies can deliver very complex systems to the enterprise market. Indeed, a business can only be built if the IT platform works flawlessly.
AMD Opteron (Socket 940), Intel Xeon Dempsey, and Xeon Woodcrest (Socket 771): Popular dual-core server processors.
Of course, you are probably familiar with the Athlon, Celeron, Core 2, and Sempron processor lines, which are desktop processors for home and office computers. But AMD and Intel have products aimed at professional customers: AMD Opteron, Intel Xeon and Itanium. The Opteron is built on the AMD64 architecture, like the Athlon 64 and Sempron processors, while the Xeon is built on the Core 2 or Pentium NetBurst architecture, depending on the model.
Professional processors typically have more interfaces—multiple HyperTransport links on the Opteron, two independent FSBs (one per processor) in the Intel world—and a richer set of features often required by server applications and workstation software.
There are two different versions of Opteron processors on the market, one using Socket 940 with DDR memory, the other using Socket 1207 (Socket F) and DDR2 RAM. As with all AMD64 processors, the memory controller is part of the processor, which can be a significant advantage as the number of processors grows: not only will you get more memory controllers to install more memory, but each processor will work with its own block of memory. Of course, in this case, coherence problems arise and the complexity of multiprocessor systems increases, but the overall throughput also turns out to be higher. Opterons for Socket 940 Opterons use PGA packaging, that is, the legs are on the processor. Opterons for Socket 1207 switched to LGA packaging, with the pins on the socket and the flat pins on the processor.
Dual-core processors should be chosen these days. Dual-core processors, albeit with a lower clock speed, outperform single-core models in the server market. Dual-core Opterons for Socket 940 are built on the cores of Egypt and Italy, the latter version is more advanced. But today we recommend choosing models for Socket 1207 (Socket F), thanks to the support for DDR2 memory and the ability to upgrade to quad-core processors, which will appear sometime this year.
The current AMD Socket F with 1207 pins is suitable for current dual-core and future quad-core Opteron processors.
Intel Xeon processors are available in different forms, with previous versions using Socket 604. Modern platforms are based on Socket 771, which refers to LGA sockets. There are different Intel Xeon processors, but we recommend sticking with dual-core models only. The table http://www.intel.com/products/processor_number/chart/xeon.htm has a complete list of processors.
Models 5030 to 5080 are manufactured on a 90nm process and are based on the now outdated NetBurst architecture. We recommend Woodcrest-based Xeon processors with model numbers ranging from 5110 (1.6 GHz) to 5160 (3.0 GHz). They are manufactured using 65nm technology, require less power, but provide high performance. The E53xx line is built on quad-core processors clovertown with frequencies from 1.6 to 2.66 GHz.
Xeon processors do not have an integrated memory controller. Instead, they rely on the motherboard chipset's quad-channel DDR2-667 memory controller. To ensure sufficient throughput for dual- or quad-core processors, modern platform Socket 771 (Blackford) provides two independent FSBs (DIBs), one for each processor.
Intel is the first manufacturer to introduce quad-core processors. Clovertown is assembled from two dual-core Woodcrest crystals placed in one package.
Intel Xeon Dempsey (65nm NetBurst), Woodcrest (65nm dual-core Core 2) and Clovertown (65nm quad-core Core 2).
Server memory works on the same principle as conventional consumer PC memory. The current standard is DDR2 (Double Data Rate SDRAM second generation). DDR2 works with more prefetch buffers (4 instead of 2), so the interface frequency can be doubled compared to DDR1.
Compared to consumer memory, professional memory has two different mechanisms designed to preserve data integrity. The register memory contains a small chip, the so-called "register", which is responsible for updating the signal. If the memory of a conventional PC cannot consist of more than four (or sometimes six) DIMMs - the signals pass through all the memory modules and attenuate, then registered memory easily allows you to install eight modules. In addition to the register, DDR2 memory contains an on-chip termination that prevents signal reflection.
The second mechanism is the ECC error correction code. Instead of storing the standard 64 bits per channel, ECC DIMMs add another memory chip that can store 8 more bits, allowing data to be recovered. Therefore, single-bit errors can be corrected on the fly.
All AMD Opteron processors for Socket 940 require DDR333/DDR400 registered memory, while the Socket F generation (Socket 1207) requires DDR2-667 registered memory.
Fully-Buffered DIMMs (FB-DIMMs) use a so-called buffer component, a high-power chip that converts parallel signals to a serial interface. Its main purpose is to connect more than eight memory modules per controller. With a quad-channel Intel DDR2 memory controller, you can install eight 2GB DIMMs on each of the four channels if motherboard manufacturers want to support this configuration, of course.
FB-DIMMs are more expensive, run hotter, and run no faster than conventional registered memory. Yes, they are most likely the future of servers with large amounts of memory, the same technology is used for current Intel Xeon platforms.
Click on the picture to enlarge.
As an example, we took the Asus P5MT server motherboard (it is used in entry-level servers, since it allows you to use conventional processors, rather than more expensive server ones). Server motherboards do not support overclocking and are usually equipped with a large number of interfaces, as well as expansion slots with high bandwidth.
The 133MHz PCI-X bus continues to be the dominant interface for expansion cards. It is built on the parallel PCI bus, which is found in almost every PC today. PCI-X is 64 bits wide, while the PCI bus in your computer is 32 bits wide. PCI-X 133 supports bandwidth up to 533 MB/s. However, it should be remembered that the bandwidth of the PCI-X controller is distributed among all connected devices.
The PCI Express (PCIe) interface is more modern. PCI Express is a serial interface that uses multiple lanes to connect a device to a controller. Professional expansion cards use PCIe x4 slots (four lanes), but there are also x1, x8 and x16 PCIe cards/slots. PCIe x16 is commonly used for high performance graphics cards, graphics workstations carry two full PCIe x16 slots for two graphics cards.
Motherboards for servers and workstations usually contain an integrated network controller. It can be built on the same components found in consumer-grade motherboards, but usually more powerful chips are built in to provide, for example, hardware support for TCP/IP computing or other features to increase performance.
This board is equipped with four DDR2 memory slots, one Socket 775 socket for installing a Pentium 4 or Core 2 processor, one 32-bit PCI slot, one PCI Express x16 slot for a video card or a powerful storage controller, and two PCI-X 133 slots. Broadcom gigabit Ethernet controllers are responsible for the networking capabilities. An ATi GPU is installed on the motherboard. It is, of course, outdated, but it is enough to display the desktop or command line, which is what is required for server OSes.
All other interfaces and components can also be found on consumer hardware: southbridge, UltraATA/100 or Serial ATA controllers, voltage regulators, etc. The big difference, again, is the validation process, during which manufacturers test their products against others and publish compatibility lists.
The ATi RageXL chip is years old and doesn't support 3D graphics, but it's good enough for servers. Moreover, there most of the time no one is looking at the screen.
A little higher, we already mentioned the motherboard with an integrated video card. All server motherboards are equipped with a very simple GPU with a small amount of dedicated memory - solutions that take memory from the RAM are not popular here. Today's successor to the RageXL is the ATi ES1000 GPU, which originally worked in the consumer market, but then appeared in servers due to improvements in hardware and drivers. Administrators don't even have to think about installing a custom or updated driver version: the driver comes with the OS and is certified.
Workstations, on the other hand, require more powerful hardware. ATi is positioning FireGL graphics accelerators based on the Radeon X1000 line for this market. Nvidia offers a Quadro FX line very close to the GeForce 7000 family. The difference between consumer and professional chips can be small, for example in driver optimization. Professional graphics cards provide excellent performance in specialized applications, but they also cost a lot more.
Hard drives are another interesting aspect of servers and workstations. A few years ago, server hard drives used the Small Computer System Interface (SCSI) interface and 10,000 or 15,000 rpm spindle speeds, which were noticeably faster than 7,200 rpm desktop drives. Server hard drives are still faster, although the difference isn't that big anymore.
The professional hard drive market is divided into three segments. The first high-capacity segment uses conventional 3.5" Serial ATA hard drives validated for 24/7 operation. The performance segment is trying to maximize storage density, so we are seeing more and more 2.5" high-performance hard drives for 10 000 rpm with Serial Attached SCSI (SAS) interface. The high-end segment relies on 15,000 rpm SCSI or SAS hard drives.
Hard drives for servers and workstations typically require active cooling as they are optimized for maximum reliability and performance. All professional hard drives come with a five-year warranty.
Power supplies for the professional sector are specially designed with maximum reliability in mind. Any decent power supply can fix the effects of one missing phase, but professional solutions can deal with more serious failures. Some also provide surge protection, although here we get an overlap with the area that lies in the area of responsibility of uninterruptible power systems (UPS).
Professional power supplies are modular and provide redundancy in the form of two modules, each of which is able to provide sufficient power to the system. If one power supply fails, the system will continue its work from the second unit.
Question about difference between a server and a regular computer arises for any programmer or developer: sometimes as a simple interest, sometimes as a practical task. It is a pity that many managers do not know the difference when trying to organize complex enterprise-level management systems based on office PCs. And after that, they wonder for a long time why something “wrong” works.
A server is, first of all, a network computer whose task is to allocate resources for ordinary computers in its network. If the network is small, then a regular PC can also be a server. There is no difference between computers here, but there is a difference in the software - the server uses the server version of the operating system, as well as additional services and programs, which are also called servers: mail, web, DHCP, etc. With the growth of the network, the power of the server should increase proportionally, and that is why you have to look for stores selling server equipment. And you will definitely need it:
- More powerful hulls. Servers have significantly larger motherboards due to the excess of connected interfaces and the number of processors.
- More power supplies. Often 2-3 PSUs are used, and their hot swapping is allowed. In general, server cases and power supplies are often placed in special racks, and “regular” plug-in units can dramatically increase server scalability.
- High-speed network equipment. It is in the vicinity of the servers that the fastest cables and other interfaces are laid.
- Hard drives, memory. Server programs are very voracious in terms of resource consumption, so the disk memory here is measured in tens and hundreds of terabytes, and the operational memory is 32-64 or more gigabytes. Moreover, for servers, RAM is produced with error control - ECC, and it is unsuitable for a PC.
In general, server appetites grow depending on many factors. The price per server increases much faster, so servers are often not bought, but rented. Moreover, not every enterprise can afford to maintain professional staff for its setup and round-the-clock support, as well as the maintenance of a special room - a server room, where an ideal microclimate for equipment should be provided.
Interestingly, the desire of many gamers to “play on the server” is unrealistic, since graphics are the key factor in games, and in servers, graphics, including monitors, are an unnecessary thing, and are used only to control the state of the system. So players will have to settle for regular PCs with 2-3 processors, while server easily uses hundreds of processors.
--------
A server is a powerful computer serving other computers on the local network. ComputerBild will tell you how a server differs from a regular PC.
In any computer network of any size, there is a constant need to share resources between multiple computers, whether it be a shared Internet connection, access to multimedia files, or printing documents on a single printer. The computer that provides these resources to other machines is called a server. The nature of these resources determines the type of server. The file server stores data, the print server receives documents and sends them to a printer connected to it, connecting to a proxy server to access the Internet, computers share an access channel ... These and other functions can be performed by both different machines and one computer.
The difference between a server and a regular PC
Servers that are used in home "locals" and in small businesses, as a rule, differ from ordinary PCs only in the software installed on them. Another thing is the servers of large organizations. The load on their computing resources and storage devices is very high. These machines must accommodate large volumes of documents and provide high speed access to them. Also, no less important, the server requires uninterrupted operation and high fault tolerance. Therefore, large servers, in general, consist of more complex and high-performance hardware than ordinary PCs. At the same time, some hardware components, the functions of which are secondary to the server, turn out to be weaker than their counterparts in the home PC. These are the components that distinguish servers from simple computers.
Powerful processors. Servers use special CPUs, such as Intel's Xeon or AMD's Opteron. More exotic "stones" are also used, such as Intel Itanium. Entry-level servers, like simple PCs, have one processor, larger ones - from two to eight. The most powerful servers are clusters of hundreds of processors with accompanying "strapping" - motherboards, drives, etc.
Large amount of RAM. If a home PC needs a couple of gigabytes of “RAM” for full-fledged work, in the worst cases - 4, then a powerful server needs 8-16 GB and even more. The server memory modules themselves, as a rule, have an error correction function - ECC (Error Correction Code). Due to this, errors in writing and reading data caused by a malfunction of the electronics or a defect in the memory chips will not lead to interruptions in the operation of the "software" or the system freezing, as would happen with a conventional computer.
Capacious drives. Most servers store data on fast, high-capacity hard drives that are combined into RAID arrays. Since the server simultaneously requires high speed and fault tolerance, array formats combine data sharing across multiple hard drives with duplication of information on other "screws". Often there is also the ability to disconnect and connect drives "hot" - i.e. without interrupting the system.
Broadband network connection. If the server is used to share the Internet with dozens of computers, then it connects to the global network using a "thick" channel with high bandwidth. For this, fiber optic lines or radio channels are used. Client computers connect to the server using familiar Ethernet or Wi-Fi technologies.
The hardware that performs secondary service functions is also different for servers from the periphery of client PCs.
Frame. The tasks of the server determine its design. Entry-level servers look like regular PCs, only scaled up to fit a larger motherboard and storage array. Servers have more powerful cases as rack cabinets, and often they consist of several units (computers, routers, etc.) in separate cases. A very powerful server cluster may consist of several dozen such cabinets. A computer enclosed in a compact rack-mounted chassis is called a thin server (blade server).
Power unit. Entry-level servers are equipped with one or two PSUs. As the number of functional blocks of the server increases and their “appetites” increase, the number and power of power supplies increase. Often there is the possibility of "hot" replacement or connection of an additional PSU.
output devices. Since the output of video and sound is among the client, and not the server tasks, the devices for this purpose are either absent from the servers at all (then the system is controlled remotely from the client PC), or they are rather primitive.
Features of the work and hardware of a powerful server dictate the need for its specific maintenance.
Uninterrupted power supply. Industrial surge protectors and uninterruptible power supplies prevent data loss in the event of "failures" of voltage in the electrical network, and also reduce the likelihood of damage to "iron" and files during its sudden drops. In some cases, when uninterrupted operation is required in spite of everything, backup power generators are used.
Enhanced cooling. Most servers, like client PCs, are air-cooled. The problem of increasing heat dissipation along with performance is solved with the help of enhanced ventilation of cases and rooms where servers are installed. Servers do not require enhanced cooling of individual components, so they do not have water cooling systems
Specialized software. Servers run operating systems adapted to server tasks, such as Windows Server 2003, special versions of Linux or FreeBSD, or pure server operating systems such as Solaris by Sun Microsystems. Server programs are used as the main software, such as Apache for maintaining websites or Microsoft Exchange Server for receiving and sending e-mail.
Server types
A server is not only a computer, but also software that manages shared resources and access to them. Several server programs can run simultaneously on one computer. In everyday life, speaking, for example, of a "mail server", they mean a combination of "hardware" and "software". Depending on the functions that the software performs, several types of servers can be distinguished. All of them can be divided into two groups: servers, whose task is to store data and provide access to them to users, and servers that control the transport of data in the network and support its operation. The first group includes the following types of servers.
File server. Its tasks include storing files and providing access to them by client PCs, for example, via FTP. File server resources can be either open to all computers on the network, or protected by an identification system and access rights.
Media servers are a type of file server. They are designed to store photos, music, movies and other multimedia content. It is not necessary to use a computer as such a server. You can buy a NAS device or even get by with a compact external hard drive that connects to your network via Ethernet or Wi-Fi.
The print server receives print requests from computers on the local network and sends them to one or more printers connected to it.
game servers. Computer game developers open special servers where users can play with each other. Once upon a time, 3D shooter and strategy servers were the most popular, allowing only one match at a time or several at the same time. A rare house or quarter “local” does without such a server. Nowadays, servers of various MMORPGs (Massive Multiplayer Online Role Playing Game) are more in demand, on which hundreds and thousands of people can play at the same time (for example: Lineage 2 and World of Warcraft games).
Email servers. An e-mail cannot be sent directly to the recipient - first it goes to the server where the sender's account is registered. The latter, in turn, sends a "package" to the recipient's server, from which the latter picks up the message. Despite the fact that both receiving and sending letters are performed by the same server program, formally these functions are assigned to different servers with different addresses.
Instant messaging servers. Messenger programs - AIM, ICQ or MSN Messenger - work with the help of a network of communication servers, operating on the same general principle as mail servers.
web servers. These servers provide access to web pages and related resources such as pictures. Sites with high traffic or advanced functionality are hosted on several servers at once.
Data servers store various kinds of materials necessary for the functioning of servers for other purposes. For example, some of the text, graphics, and style elements of a website may be hosted on a separate data server. When the user opens the home page of the site, the web server sends a request to the data server to receive the necessary materials. The database server searches for the requested data and sends it to the web server. It, in turn, generates a web page and sends it to the client computer.
The list of servers that manage traffic transport includes the following varieties.
DHCP servers. Dynamic Host Configuration Protocol provides automatic distribution of IP addresses between computers on a network. This technology is widely used in local networks with general access to the Internet.
DNS servers. The function of a DNS server is to translate the domain names of servers into IP addresses. Data transmission in networks is carried out using IP addresses, a site with an unchanged domain name can “move” from one server to another more than once, changing its IP address. Therefore, the correspondence tables of IP addresses and domain names in the DNS (Domain Name System) system are regularly updated, and the servers synchronize them with each other.
Proxy servers act as intermediaries in the transmission of data over the network - from computer to computer. They are used either to organize shared access to the Internet, when traffic control and filtering is required, or to hide the IP address of the PC from the “interlocutor” computer, because. the latter, when communicating through a proxy server, will “see” only the proxy address.
cache servers. So that each time you open a web page, the client computer does not have to re-request all the data that makes it up, intermediate drives are used - cache servers. If the page requested by the user has not changed since the last request, then it can be loaded not from the "native" storage, but from the bowels of the cache server.
Client-Server and Peer-to-Peer Architectures
If the computer that provides resources is a server, then the computer that uses them is called a client. In addition, the client, like the server, is also a program for accessing resources (for example, an email client or an instant messaging program).
Most traditional Internet services are based on the client-server architecture. But recently a fundamentally different network organization has become widespread.
In a Peer-to-Peer (P2P) architecture, all computers are equal and each store a portion of the total amount of data. In this case, each machine acts both as a client and as a server. The most successful example of P2P implementation is file-sharing networks (eDon-key2000, Bit-Torrent). When you receive a file from such a network, you simultaneously download its fragments from dozens of computers. Due to the distribution of data, peer-to-peer (otherwise peer-to-peer, decentralized) networks are characterized by high fault tolerance and speed.
In fairness, it must be admitted that most peer-to-peer networks cannot do without servers at all. For example, file-sharing networks use servers (trackers) to manage traffic and account for it.
home server
All modern operating systems have server capabilities. With their help, you can allow users of other PCs to access data on a hard drive or a printer connected to a computer, as well as "share" an Internet connection. In addition, a home server can be used for backup data storage or, by making it available via the Internet, work with documents on it from any PC connected to the global network.
“Rising” a home server for storing files and sharing access to the Internet is not as difficult as it might seem to an inexperienced user. To do this, you will need the following components.
Computer. For a file or simple web server, a computer with a processor no weaker than a Pentium II or Athlon, 256 MB of RAM and a CD-ROM drive is enough. If you plan to run a game server on your computer (a very popular initiative in small local networks), you will need a more powerful machine.
As a first step, you can run a Linux server on it from the Live-CD. After playing around with it, you will decide whether you should move on to more serious activities. If you still decide to use your computer as a server all the time, then Linux will need to be installed on your hard drive. For this, 10 GB of free space is enough. The rest of the space will be left for files and additional software (primarily server programs).
It's not a bad idea to use an old laptop as a server. This will save you money on your energy bills over a long period of time. In addition, the folded laptop takes up very little space. The only drawback of the laptop in this case is the limited possibilities for connecting drives.
Linux distribution. With a free version of Linux (Open SuSe, Ubuntu or Knoppix) you can create a server that has all the features you need for home use.
Most Linux distributions also have paid versions - for example, in the case of SuSe, it is called SuSe Enterprise Server. This version of Linux is distinguished by additional technical support from the manufacturer and an extended set of programs.
/>
Network hard drives
DimonVideo DimonVideo
/>
WLAN routers and network hard drives
The operation of a home network rarely requires the use of a separate PC as a server. Depending on the amount of data it will store and the tasks it will perform, there are two cheaper options to choose from.
WLAN router plus external hard drive
Many wireless routers have a USB port where you can connect an external hard drive. Any computer on the network will have access to the data stored on it.
Network hard drives
Network Attached Storage (NAS) is a compact and inexpensive (compared to a separate PC) server that performs only one function - data storage. Powerful NAS have a rich set of interfaces and the possibility of remote configuration via a web interface (similar to routers). Simpler options for such a solution are ordinary external hard drives with an Ethernet or Wi-Fi network interface.
It makes sense to install a server based on a full-fledged computer only if the NAS functionality is no longer enough: for example, you need to “raise” a small game server on your home network or a website. An old Linux computer will suffice for this purpose, although Windows can also be used.
">Hello friends!
Today we will talk about serious "pieces of iron" arranged more abruptly than a household computer.
Hardware and software server
Many of us have heard the buzzwords “server” and “file server” and the phrases “mail server”, “dedicated server”, etc.
In this article, we will take a look at what is a server how it works and how it differs from a home or office computer.
The word server is derived from the English term to serve (serve). So the server is the "server".
It doesn't sound very nice, but it's clear what he's doing. Serves customer requests.
You can serve at the hardware ("iron") and software levels.
Therefore, the concept of server includes two - "hardware server" and "software server".
We will immediately indicate the first difference between a server and a personal computer (PC). The user is constantly working on a personal computer. And the server (hardware), in most cases, stands quietly in a separate room and works independently.
An ordinary user does not work for him, only sometimes a system administrator sits down to him - for configuration or other service manipulations. It is necessarily included in the local network (otherwise, how will it serve client requests?). Therefore, it can be accessed from any computer from this network (with appropriate settings, of course). A hardware server, unlike a PC, operates 24 hours a day, seven days a week.
Software server
A software server is a software package that serves client requests.
This ideology implies two parts software package - server and client. The main part is the server part. It is located on a hardware server (physically - on its hard disk).
This ideology is used by many programs, in particular, accounting. The main part of the program and the user database are stored on the server. The user interacts with the main part through the client part located on his computer.
Several (or many) clients can work in the main module at the same time. Therefore, the hardware must have the necessary processing power. Depending on the type of software server, different requirements may apply to the hardware.
Mail server is a program that works with mail. Of course, there is also its hardware part, where, in fact, letters are physically stored.
Any search engine includes a mail service. Currently, getting yourself a mailbox (or even several) on the server is not a problem.
There is such a thing as a "file server". This is a repository of many files, where many users are allowed to access. There can be a lot of files, so the hardware must have a large disk memory. The software part is, no doubt, very interesting, but our blog is about hardware, so let's take a closer look at how a hardware server works.
We note right away that the server can work for months without shutting down. Therefore he must have much more reliable Hardware. Increased reliability is provided, in particular, by better and more expensive components.
RAM
Used with ECC parity check(Error Correction Code). The fact is that there is always some non-zero probability of an error in RAM.
It's one thing when an error or crash happens on the user's computer when they're running their own application (rather than on the server). The cost of such a mistake is often small.
Another thing is when someone's transaction for a million dollars is “lost” on the server.
Errors on the server are also dealt with by hardware, using redundant data coding. The information is supplied with additional (service) bits, which makes it possible to detect and correct some (not all possible!) most probable errors.
Such memory modules, unlike conventional ones, have an odd number of chips. So, in conventional modules used in conventional or office computers, most often 8 microcircuits are installed. 9 microcircuits are installed in server modules. Server memory with the same amount and speed costs significantly more expensive ordinary.
Winchesters
Multiple users can access the server at the same time. Therefore, in addition to high reliability, the server is also required to increased performance. The servers use both conventional (electromechanical) and solid state SSD(Solid State Drive) drives.
The latter have increased performance, but they are much more expensive (for the same volume). SSD drives do not contain rotating parts. Data is written to semiconductor memory cells. We note right away that you can write data to a memory cell of an SSD drive a limited (albeit large) number of times.
To reduce data access time, servers use electromechanical hard drives with a high spindle speed - up to 15,000 rpm. Such hard drives heat up more, so the problem of heat dissipation becomes even more urgent.
Server hard drives use more intelligent interfaces - SCSI and SAS, which have more functions and more flexibility.
As a rule, hard drives in servers have the ability to hot swap(hot swop), i.e. without turning off the power.
Turning off and then turning on the server is not always a quick procedure. The server "does not like" shutdowns. Sometimes, to put it into operation, an increased consumption of nerve cells is required :-) The possibility of "hot" replacement is provided by a special design of the "pocket" for the hard drive. When the drive is pulled out, the power contacts open first, and then the information contacts. During installation, the power contacts are closed first, and then the information contacts.
Reservation
Servers always run reservatione. For this, there is a so-called RAID (Redundant Array of Independent Disks, a redundant array of hard drives). Information files are duplicated in a certain way. There can be several ways to duplicate, so there are RAID-0, -1, -2, -3, -4, -5 and -6. Now, even if some hard drive fails, the information will not be lost.
There is a hardware controller for a RAID configuration. This is what most home or office computers do not have. The RAID controller (more precisely, the part that stores the configuration) can have its own backup power source - a small capacity battery.
Cooling
The server contains at least two processors, each of which can have several cores, several hard drives, memory modules.
All this goodness consumes a lot of energy. Therefore, the server needs efficient cooling system.
Cooling is carried out with the help of radiators and , just like a regular PC. In addition to the fans installed in the power supply unit(s), several additional fans are installed. All fans have increased working resource and are made on the basis of rolling bearings. You will not find cheap "wind blowers" with plain bearings there. Fans can have different dimensions depending on the design.
The cooling system is organized in such a way that not only the processor and chipset heatsinks are blown, but also - for sure! - winchesters. All fans contain speed sensors. There are also several temperature sensors. When the temperature inside the server rises, the control system increases the fan speed for more intensive cooling.
In conclusion, we note that to increase the strength of the air flow, the fans can be duplicated.
To do this, the fans are installed close to each other and blow with a common air flow. In the event of a fan failure, the control circuit may beep or display service messages on the monitor screen.
Let's pause at this point. In the second part of the article, we will get acquainted with power supplies, the design of servers and some subtleties that not everyone knows about. Do not miss!
Victor Geronda was with you.
See you on the blog!
