Linux in large companies: why it pays off

When Linux is mentioned in a company, the first association is often “that free thing the IT people use”. Yet most internet servers run on it, as do almost all large cloud services, most phones and a great deal of industrial equipment. Large companies do not choose it out of conviction, but because it pays.

Where Linux runs

When people talk about Linux in companies, they almost always mean servers, not employees’ computers:

• web servers and business applications
• databases
• file storage and backups
• virtualisation and containers
• network equipment, firewalls and VPN
• monitoring, logs and automation

Employees keep working on the computers they know. Linux is in the background and they do not see it.

Why it pays off

A cost that does not grow with the number of servers

With commercial systems, the licence is paid per server, per processor core or per user accessing it. Every expansion means a new licence.

Linux has no per-installation fee. The tenth server costs as much as the hardware under it. For a company with a few dozen virtual servers, the difference over five years is measured in tens of thousands of euros.

That does not mean it is free. The cost is in the people who maintain it and, if wanted, in a support subscription. More on that below.

Stability

Linux servers routinely run for months without a restart. Most updates are applied without interrupting work, and those that require a restart can be planned for a time when they bother nobody.

For systems that must be available all the time (warehouse, production, web shop), that is a direct business benefit.

Security

No system is secure by itself, but Linux has several advantages:

• Clearly separated privileges. Programs and users run with the minimum rights they need.
• Fast patches. Security flaws are generally fixed within days, and updating the whole system and all software is one command from one source.
• Open code. Thousands of people review it, so flaws are harder to hide.
• A smaller attack surface. A server is installed with only what it needs, with no graphical interface and no superfluous services.

A badly configured Linux server is still vulnerable. The advantage only shows with regular maintenance.

Control and independence from a vendor

With a commercial system, the vendor decides when support for an old version ends, how the price list changes and which functions move to a more expensive package. The company adapts.

Linux does not belong to one vendor. If one distribution changes its terms, there is another, and data and applications move with relatively little work. The formats are open, so data remains readable in ten years.

Automation

Everything that can be done by hand on Linux can also be written as a script. That means twenty servers are set up the same way as one, the setup is repeatable, and it is possible to check what was changed and when. For a growing company, that is the difference between one administrator and five.

A long life

Distributions intended for companies have support of ten or more years per version. A system set up today does not have to be replaced because the vendor decided it is obsolete. Linux also runs well on older hardware, so equipment is used longer.

What is on the other side

Knowledge. Linux needs people who know it. That is the biggest real cost and the most common reason companies do not go for it. The solutions are an employee with that knowledge or an external partner who maintains the systems.

Support. Free distributions have no phone number to call at two in the morning. Companies that need that pay for a subscription to a commercial distribution or contract support from a partner.

Software that runs only on other systems. Some business software, especially older software, does not exist for Linux. That part then stays where it is, and the rest moves to Linux.

Workstations. On employees’ computers Linux is still a rarer choice, because of office software and habits. That is why the usual starting point is servers.

Where it makes the most sense

• Web and application servers. Linux is the standard here, and alternatives are the exception.
• Databases. The most widely used open-source systems run best on it.
• Storage and backups. See the article on the 3-2-1 rule.
• Virtualisation. An open-source platform for virtual servers and containers, with no per-core licence.
• Network and security. Firewalls, VPN, traffic monitoring.
• A private cloud. File sharing and collaboration on your own servers, with no per-user subscription.

How to move

Replacing everything at once is not necessary and rarely wise.

1. List the servers and what runs on them. For each, note the licence, its cost and when support ends.
2. Start with what is new. Let the next server you need be Linux, if the application allows.
3. Move what is easiest. Web, storage, backups and monitoring usually have no obstacles.
4. Leave what has to stay. A mixed environment is the normal state in most companies.
5. Agree maintenance. Who updates, who monitors and who responds when something stops.

How much can be saved

An honest calculation compares total cost over five years: licences, hardware, support and people’s work. In that calculation Linux generally comes out better the more servers there are, because licences grow with the number and knowledge does not.

For a company with one server and nobody who knows Linux, the saving may be small or none. For a company with ten or more, it is usually substantial.

The servers and networks we set up are built mostly on Linux and open source, and we use the same approach for our own systems. If you want to know what would make sense to move in your case and what it would bring, that is part of technology consulting. Get in touch.