What Are Remote Desktops? An Honest Three-Step Primer

"Remote desktop" is a phrase that gets used to describe at least four different things. When a customer calls us and says they want remote desktops, the first 15 minutes of the conversation is usually figuring out which of the four they actually mean. This is a problem worth solving because the four architectures have very different costs, different failure modes, and different users they serve well.

This is the honest three-step explanation of what a remote desktop is, what varieties exist, and how to pick between them.

Step One: Understand the Core Idea

A remote desktop is a computing session where the CPU, memory, storage, and operating system live on a host in a data center, and the user's input and display live on a device somewhere else. You are using a computer that is not in the room with you. Your keyboard and mouse generate events that travel over a network to the host. The host runs applications, updates its screen, and sends the pixels back to your device.

From your perspective the experience is supposed to feel like a local computer. From the engineering perspective it is a continuous loop of "capture input, send input over network, process on remote host, capture frame, encode, send back, decode, display" that runs dozens of times per second.

That is the whole concept. Everything else is variation on the theme.

The value proposition of a remote desktop, as we pitch it to customers, is four things:

• Data stays central. Sensitive information never lives on the endpoint. This matters for healthcare, legal, financial, and government customers for compliance reasons, and for everyone else because a lost laptop with a remote desktop on it is just a lost piece of plastic.
• Endpoints can be cheap and replaceable. A $250 thin client or an old Chromebook can run a remote desktop session as capably as a $2,000 laptop, because the heavy lifting happens on the host.
• Management is centralized. Patching, updating, and configuring 200 remote desktops is easier than patching 200 laptops because the desktops all live in one place.
• Access is universal. Any device with a compatible client can get to the same desktop environment. Home, office, branch, or on the road, it is the same session.

These benefits are real. They also come with real tradeoffs — network dependency, protocol tuning, profile management — which is why deciding that a remote desktop is the right answer has to be done per use case, not in the abstract.

Step Two: Know the Four Varieties

The phrase "remote desktop" covers at least four distinct architectures that people routinely confuse. Here are the four, with honest summaries of where each one belongs.

Session-Based Desktops (RDS / Terminal Services)

Also called Remote Desktop Services or Terminal Services. One Windows Server runs many user sessions simultaneously. Each user gets their own desktop experience but shares the underlying OS, kernel, and driver stack with everyone else on the host. Cheap, dense — you can fit 30 to 100 light users on a single Windows Server — and well-suited for call centers, data entry, and any workload where every user runs the same handful of applications.

Limits: users share a single OS instance. One user's runaway process affects everyone. Application compatibility with server-multi-user mode is sometimes a problem. Not suitable when users need administrative rights.

VDI (Virtual Desktop Infrastructure)

Each user gets a dedicated virtual machine running a client OS — typically Windows 10 or Windows 11. The VM can be persistent (your VM sticks around between logins, your changes are saved) or non-persistent (the VM resets at logoff, you get a fresh one next time). VDI is more expensive per user than RDS, because each user has their own VM, but it is more isolated and supports a wider range of applications including ones that do not work on a shared-server model.

Limits: higher cost per user, higher storage footprint, more complexity in profile management, and more sensitive to hypervisor and storage performance.

DaaS (Desktop-as-a-Service)

VDI that is run as a managed service by a provider. Your users get desktops but you do not operate the infrastructure. Azure Virtual Desktop (AVD), Windows 365, Amazon WorkSpaces, Citrix DaaS, and similar products all fit here. You pay per user per month, the provider handles the stack below the guest OS, and you handle the desktop image and the users.

Limits: less control, vendor lock-in to some extent, ongoing cost instead of capital expense, and the provider's pricing determines your cost structure.

Remote Access to a Physical PC (TeamViewer, RealVNC, Microsoft Remote Desktop to a desktop PC)

Technically a remote desktop in the literal sense, but a different architecture entirely. Instead of a data-center host, the remote endpoint is a real workstation somewhere. This is what "I left my PC at the office and need to get to it from home" looks like. It is not designed for scale — it is one-to-one, peer-to-peer, and the target machine has to be powered on and reachable.

Limits: does not scale to hundreds of users, requires the target machine to be available, and offers none of the centralization benefits of VDI or DaaS.

These four are not interchangeable. A customer asking for "remote desktops for 300 students" usually means VDI or DaaS. A customer asking for "a way to get to my office PC from home" means RDP-to-workstation. A customer with 200 call center agents on identical workflows almost certainly wants RDS. Getting the architecture right up front saves a six-figure mistake later.

Step Three: Pick the Right One for Your Actual Use Case

We walk customers through the same decision tree every time. It is short.

How many concurrent users? Under 20, any of the four architectures work, pick based on other factors. 20 to 200, lean toward RDS for uniform workloads or VDI for varied ones. Over 200, the operational complexity favors DaaS unless you have real staff to run VDI yourself.

How varied are the applications and user needs? Uniform workloads (a call center, a student computer lab, an exam room terminal) work great on RDS. Highly varied workloads where power users install software, run dev tools, or need administrative access need VDI or DaaS — the session-sharing model of RDS will not hold up.

Are the users persistent or transient? Students who sit down at a different machine every class period want non-persistent VDI where the desktop resets between users. A developer who installs tools and customizes their environment wants persistent VDI or DaaS.

Is the organization willing to run the infrastructure? Running VDI well requires storage, networking, and operational skills that smaller IT teams genuinely do not have. If the honest answer is no, go with DaaS. The premium is worth it compared to a half-working self-hosted VDI.

What is the compliance story? Regulated workloads usually work fine in DaaS if the provider signs the appropriate BAA or data-processing agreement. Extremely sensitive workloads sometimes require on-prem VDI in a sovereign data center. Know which bucket you are in before you start.

There is no single right answer. There is a right answer for your situation, and the job of a good provider is to help you find it honestly instead of selling you the thing they happen to offer. We have walked customers away from remote desktops entirely when the answer was "you should just buy everyone a laptop." That is sometimes the correct call.

Remote desktops solve a specific class of problems well. Make sure your problem is in that class before you spend the money.