What are servers?

A server is a computer that receives requests over a network, performs requested functions and sends back responses. The term can also refer to the software package that provides the service.

Any computer can act as a server, but many manufacturers offer dedicated systems designed for server roles. Such a system might have more memory and storage than a regular desktop PC, so it can process requests from many clients at once. It may use technologies such as error-correcting RAM and RAID storage, to reduce the risk of a hardware failure leading to downtime or data loss. A second, redundant power supply may be built in too.

Server systems come in two main formats: a tower server looks like a desktop PC but is larger, so it can accommodate more expansion cards and storage. Rack servers conversely use a compact format that’s designed to be mounted in a cabinet, saving space at the expense of expandability.

People have lots of questions about servers, which we have tried to address in this article. You can jump to a specific question by clicking the link.

• What’s the difference between a computer and a server?
• What can servers do?
• Why are servers so expensive?
• What are cloud servers?
• How much does a cloud server cost?
• Who needs servers?
• Do I need a server?
• Can I build my own server?
• How many servers does a company need?
• Can a server provide multiple services at once?
• Where are the servers for the internet?
• What is the future of servers?

What’s the difference between a computer and a server?

There isn’t necessarily any difference. The term “server” describes a role rather than a specific piece of technology, so any computer can be a server. Often though the term “server” is used specifically to refer to a business-focused computer system that’s designed to prioritise reliability and performance, so it can serve multiple users at once over a network.

What can servers do?

There are many types of server. A medium-sized company might operate email and storage servers, along with a database server that allows staff to find and cross-reference customer data as needed. Businesses with more advanced needs might run a CRM server, or some other type of application server that works in partnership with a front-end client application, allowing users to collect, manage and analyse data.

Servers also underpin the network itself. DHCP and DNS servers handle addressing requests, enabling computers to connect to one another, either internally or over the internet, while web servers receive HTTP requests, and send HTML pages in response.

Why are servers so expensive?

While a regular desktop PC is normally used by one person at a time, a server needs to handle dozens or hundreds of users at once. It therefore needs more memory and storage than a typical personal computer, and often a more powerful processor.

What’s more, server hardware is expected to be working all day every day, possibly for years on end – so reliable, high-quality components are a must. A server might use technologies such as error-correcting RAM and RAID storage, to reduce the risk of a hardware failure leading to downtime or data loss. A second, redundant power supply may be built in too.

What are cloud servers?

Cloud servers are those that offer services over the internet, especially ones that store or process data. The details of the actual server hardware and operating system are normally concealed from the user: they just interact with the service, while the technical side of things is taken care of by the service provider.

How much does a cloud server cost?

Leasing capacity on a cloud server can be very cheap. Providers such as Amazon and Microsoft offer free access for lightweight workloads, while paid-for remote computing and storage services start at just a few cents per hour of use.

The actual server hardware however can be very expensive. An individual system might not cost any more than a regular business server – perhaps a few thousand pounds per unit. However, to serve a huge worldwide user base, cloud providers will have hundreds or thousands of these servers working together in huge data centres. Microsoft recently announced plans to invest $50bn a year into growing its cloud capacity.

Who needs servers?

Anyone who wants their services to be accessible over a network needs a server. That might mean websites, search engines, messaging platforms and retail sites like Amazon, or it might mean local IT services such as central storage and email provision inside a business.

Do I need a server?

Most individuals don’t need their own servers: their needs are met by third-party services. Your communications, cloud storage, TV streaming and online shopping are likely all handled by internet servers, operated by external companies.

However, it’s perfectly possible to run all sorts of servers on your home network. For example, you could set up a personal storage server, keeping all your data in one place so it’s conveniently accessible and easy to back up. Or, you could deploy a media server, enabling you to stream your own videos and music files to any device that’s on your local network.

You can also rent access to third-party servers for a variety of purposes. For example, you might pay a web-hosting service to host your personal web page and make it available over the internet. Or, you could use a hosted computing service such as Microsoft Azure or Amazon Web Services, to run programs remotely on high-performance hardware.

Can I build my own server?

You absolutely can. You can use standard PC components, or buy specialist server parts and assemble them yourself – and you can choose whether to install a free Linux distribution, or pay for a commercial OS such as Windows Server 2022.

However, if you build your own server you won’t get the benefit of professional support. That’s a key reason why most companies buy big-brand servers: if something goes wrong, they want the reassurance that an expert can fix it quickly, or that any parts that fail can be quickly and reliably replaced.

How many servers does a company need?

This depends on the size and requirements of the business, but it’s likely fewer than you think. One physical machine can many different services – so, for example, a single server could handle email, storage and database duties for a small business.

This is often achieved using virtualisation, so that each service runs inside its own isolated environment, on its own operating system. It’s even possible to have multiple installations of Windows, Linux and other platforms coexisting on a single physical machine. As well as providing flexibility, this makes it easy to back up and manage critical servers, and helps keep any technical issues or security threats isolated to individual OS instances.

Can a server provide multiple services at once?

There’s no technical reason why a single server can’t provide any number of services, but resource management becomes crucial when multiple services are sharing memory, processor power, storage and network bandwidth. Get it wrong and slowdowns or crashes are inevitable.

If multiple services are all running under one operating system, there’s also a risk that their operations could interfere with one another – and a single software failure could take down all the services at once. Security is a concern too, as a vulnerability in one service could allow an attacker to gain access to all the software or data on that machine.

For these reasons, when running multiple services on a single physical server, it’s common to use virtualisation to run each in its own isolated environment. This significantly increases the resource requirements, but it provides better stability and security.

Where are the servers for the internet?

It’s estimated that there are around 100 million servers on the internet, providing every sort of online service under the sun. They’re spread out all around the world – but the country with the most server sites is the USA, followed by Germany, the UK, China, the Netherlands and Australia.

Many of the largest companies (such as Google and Meta) operate dozens of interconnected server sites situated around the world, so when you visit Gmail or Facebook you’re directed to the server nearest you.

What is the future of servers?

The past 20 years have seen many companies decommission their servers and switch to cloud computing services. This can bring considerable savings – and additional flexibility, as cloud resources can easily be expanded and reduced to follow demand.

However, accessing cloud services is much slower than using a local network connection, and introduces concerns over security and data sovereignty. Many companies now adopt a “hybrid cloud” approach, using both hosted services and in-house servers for the roles they’re best suited to – so while server hardware isn’t likely to die out, it’s expected to become more niche and specialised.

Was this explainer helpful?

If so, you might want to check out others we’ve published in our hardware section, including What is Intel vPro?, What are virtual machines? and What is Wi-Fi 7?

Updates

We updated this article on 26 January 2024 with fresh images, answers to several more questions and a table of contents.