Thin Client

A normal office computer handles its tasks by itself. It has a fast processor, plenty of memory, and a hard drive with all the programs. A thin client is deliberately the opposite. It is a small, low-power device with a screen, keyboard, and network connection that computes almost nothing on its own. Instead, the programs run on a powerful central computer shared by many workstations. The device on the desk merely transmits the image and reports back what the user types or clicks.

Why companies and schools rely on it

The most important reason is administrative effort. In a company with 500 workstations, an update would otherwise have to be installed 500 times. With thin clients, it is installed once on the central computer and immediately applies to everyone. Even a broken device is no drama: you swap it out, plug in the network cable, and the user continues working. Their files were never stored on the device, but always centrally.

A second reason is security. If no data is stored on the device itself, a stolen thin client is worthless. In hospitals, banks, or government agencies, this is a strong argument. In addition, it is possible to precisely define centrally who is allowed to see which programs.

Then there are costs and power consumption. A thin client often needs under ten watts, while a fully equipped PC needs many times that. The devices usually have no moving parts like fans and therefore last for many years. The catch: if the central computer or the network fails, the entire operation grinds to a halt. With individual PCs, only one workstation would be affected.

What goes back and forth between the device and the data center

Technically, a separate, isolated work environment runs on the central server for each user. This environment is called a virtual machine: a computer simulated by software that behaves like a real one. The server can run dozens of such simulated computers simultaneously. Each user sees their own desktop and notices nothing of the others.

Only the screen image is then transmitted, usually heavily compressed. Special transmission rules, called protocols, exist for this, with names like RDP or Citrix HDX. Cleverly, they do not constantly send the entire image, but only the areas that have changed. In the opposite direction, keystrokes and mouse movements flow back. Because this is little data, a modest network connection is often sufficient.

Limits become apparent with video and 3D graphics. A film or a design program changes the entire image constantly, which generates a lot of data traffic. Even small delays then become unpleasantly noticeable. For word processing, spreadsheets, and databases, however, this plays no role.

From the bank counter to the cloud

Classic thin clients are found wherever many people perform similar tasks. Checkout counters in supermarkets, counters at banks and insurance companies, reception desks in clinics, computer labs in schools. You can recognize them by the small black box behind the monitor, often no bigger than a book.

The idea has also migrated into new forms. A Chromebook, for example, stores little locally and works mainly in the browser with services from the internet. Anyone who writes with an AI like ChatGPT through the browser is using a similar pattern: the heavy computing happens in a distant data center, while one’s own device merely displays the result. In trade articles, one therefore reads that the thin client principle has returned via the cloud.

A common misconception is that a thin client is simply an old, slow PC. That is not true. It is deliberately built weak, because performance is not needed at this point. The counterpart term is fat client or rich client, referring to the ordinary computer that handles everything itself.