What Is a Zero Client?

A zero client is the thinnest kind of desktop endpoint. It has a processor or decoder chip, network port, display outputs and USB ports, but no general-purpose operating system for the user to see, no local applications and nowhere to keep user files. Everything the user works with runs on a server, as a virtual desktop or a remote desktop session.

When it powers on, a zero client loads its firmware, finds a connection broker on the network and opens a session. From then on it decodes the screen image sent by the server and sends keyboard, mouse and USB activity back. Power it off and nothing is left on the device, which is why zero clients are called stateless.

Management is central. Administrators set up and update every zero client from one console, and replacing a failed device is a matter of plugging in a new one. The trade-off is that a zero client speaks only the protocols its firmware supports and can do nothing without the server and network behind it.

The term is used for a few different designs. They share the stateless idea but differ in how they get there.

A thin client keeps a small local operating system, such as a Linux-based thin OS or Windows IoT Enterprise, so it can run several connection clients, a local browser or a few local applications. A zero client strips that away. The difference matters for security, management and flexibility.

In practice the line has blurred. Many devices sold as zero clients run a hidden, read-only OS, and many thin clients can be locked down until they behave like zero clients. Ask which protocols and brokers a device supports, and how its firmware is updated, rather than relying on the name.

Zero clients suit organisations that already run their desktops and applications centrally and want endpoints that need little attention. Their strengths are security, because no data rests on the device and there is almost nothing to attack; easy management, because firmware is the only thing to update; low power use, often under 15 W; and a long working life, because the device does no heavy processing.

The limits come from the same design. A zero client is only as good as its network and server, so a network fault stops every terminal on it. It is tied to the protocols in its firmware, and when a protocol or broker is retired, the devices may need to be replaced. Demanding graphics, video calls and unusual USB peripherals depend on the protocol’s support for them.

Zero clients and network-booted terminals are common in industrial automation. Operator stations connect to central HMI and SCADA servers, so applications are installed and updated once rather than on every machine. That model, and the split between HMI and SCADA roles, keeps plant-floor software consistent across shifts and lines. Control rooms use multi-display zero clients to view several systems from one desk.

Factory and outdoor sites need endpoint hardware built for them: fanless, sealed against dust and moisture, and able to run across a wide temperature range. Industrial mini PCs and panel PCs can work as zero clients, either by network booting or with a locked-down Windows IoT Enterprise or Linux image that opens straight into the remote session.

We supply industrial mini PCs and panel PCs for zero-client and network-boot roles, set up with the BIOS boot order, locked-down image or remote desktop client your servers expect as standard.