Ok, I'm over Android thin clients, but Windows RT thin clients just might work!

As noted below by Martin Sheppard, what follows is not actually feasible. The only way you get the ability to access a VDI desktop is if the Windows RT device is a companion to a device that already has SA.

As noted below by Martin Sheppard, what follows is not actually feasible. The only way you get the ability to access a VDI desktop is if the Windows RT device is a companion to a device that already has SA. So yes, this would work, as long as they had a desktop sitting under their desk that had SA on it. I guess it could be off to save on management, but that doesn't help the cost model :). Thanks Martin, for the info (and for raining on my parade!). This is a nice thought exercise, though, so I still encourage you to read on.

One of the best things about twitter for me is the fact that article ideas can come from anywhere. Today's article is 100% based on ideas thrown around by myself, Gunnar Berger (@gunnarwb), David Stafford (@dstafford), Nathan Coutinho (@nathancoutinho), and others stemming from a tweet I mindlessly tossed out to the world regarding the realization that you're only entitled to use Windows RT devices (like Surface RT) to access VDI desktops without buying additional licenses if the device is corporate-owned only. Employee-owned devices still require a VDA or CDL license.



After griping about Microsoft screwing people in a BYOD situation (as if consumers were actually buying Windows RT devices), the conversation turned to Windows RT-based thin clients. David Stafford suggested it along the same vein as my Android thin client on-again/off-again relationship. Dismissive at first, I began to think about what a Windows RT-based thin client might look like and why a company would go that route. Let's look at some interesting points.


While the lack of a built-in license might be troublesome for BYOD situations, few employees are buying their own thin clients. Using Windows RT as a thin client OS on a corporate-owned device would be an easy  way to avoid buying VDA licenses for thin clients. Yes, there is a cost for Windows RT, but according this article from Engadget, that cost is somewhere around $85. Keep in mind, too, that it's a one-time cost, as opposed to VDA which is $99/year. So, even if the thin client costs more, it pays for itself in less than a year. We're already saving money…that's a good start! Of course, we don't know how much the hardware would cost, but it would have to be less than a tablet, right?

Interface and applications

This is, to me, a distinct advantage of other non-traditional thin client OSes. I've written lately about how I've fallen out of love with the idea of Android thin clients. The primary reason for this is that Android isn't made for a desktop form factor, and neither are the applications. Keyboard and mouse support is bolted on rather than integrated, and even if it were perfect, the applications would also have to be written not only for non-touch inputs but also for desktop screen resolutions and aspect ratios. 

With WinRT, all of that dirty work is done. It's made for use with a keyboard and mouse (both buttons, even!), and while it can at times be atrocious to use without touch, in a limited use case such as a thin client it could be more than useable. Add a touch screen and it becomes even better.

The other bonus with regards to the interface is that the applications, every single one of them, are written to work on Windows RT and its many devices. That means that if Windows Apps catch on, those apps can be used, even deployed, to Windows RT thin clients to run alongside published applications or full-on VDI desktops. (Granted, for RDSH applications, you'd still need an RDS CAL.) In theory, someone could probably even work up some reverse seamless windows solution for use with it that could combine everything into one interface.

The last thing is Office, which, even in its limited form, might be good enough for use in a thin client scenario. It's certainly more useable with a keyboard and mouse, so it stands to reason that the possible use cases expand in this scenario, too.


This is where things get a little dicey, because there isn't a built-in way to manage Windows RT devices like there would be for managing thin clients with embedded versions of x86 Windows. Still, Windows RT devices are manageable via EMM solutions, so it's not outside the realm of possibility that thin clients would be as manageable as other non-Windows thin clients. Then again, perhaps we don't need to manage them. Brian and Jack might argue that you should treat all devices as insecure no matter what. I'm not quite as black and white on the matter, but if you are then management isn't going to be a problem.


Adding in other capabilities as they come along (Lync support?) only serves to expand the use case even further, and the fact that you can save $100/device/year on licensing means that there are more funds available to put into a management platform (that you need anyway for all those mobile devices you're dealing with). Ultimately, you could wind up actually saving money and getting more flexibility (because of something Microsoft did, no less)!

So what am I missing? Why isn't anyone making these? I'm in the honeymoon phase of this idea, so everything is wine and roses at this point. Someone hit me with a dose of reality!

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Hi Gabe,

This is something ill be targeting with ThinKiosk. If Microsoft open up RT to run on thin client form factors you'll get all of the standard benefits of windows redirection technologies and even potentially a touch interface that can be passed through to win 8 / 2012 hosted desktops.

In the thinkiosk stack a management interface is under dev and will be possible to manage the device and keep it up to date. Delivery might be a challenge with the store requirement, but we all love a good challenge!

I'd expect w release candidate available to the community mid q2.



I think this is an interesting concept. As virtual environments start to play a larger role in businesses I think we will see a lot more creativity by IT departments to cut down on licensing costs. However, as you mentioned, a lack of management capabilities may put a stop to this.  Also, pricing of the hardware may play a role as some companies may roll out customized Thin Clients from waterfalled PCs. Nonetheless,  it will be interesting to see how this scenario plays out.


How would Windows RT be any better then Windows Embedded? RT adds a much unnecessary bloat and last I checked most thin clients still only come with a couple GB of flash storage. Good luck fitting RT on that.


@Things : For starters, it would help eliminate a recurring cost (SA to get ERR else VDA).  Next it would be built for for lower cost ARM-based processors.  OEM's should be able to get higher end media capable SoC's around $20.  You mention bloat,,, OEM's should be able to tune their distro the same or better than WinEmbedded, though I'm not sure if a write file/volume filter driver exists as it does for Embedded or WinTPC.


I think you are forgetting that the free access to VDI from Windows RT devices only applies if they are "Companion Devices", i.e. the user operating the Windows RT device is also the primary user of a device covered with Windows Client SA or VDA.

Even so, it might be quite a good idea in the right circumstances. For one, if your users all have their own PC, but need a second thin client then using Windows RT on it really would give you free access to VDI that you wouldn't otherwise have and it certainly wouldn't be a bad OS for a thin client, but it does seem like you need to have somewhat unusual circumstances for this to beat Windows Embedded as the preferred OS.

I've included the relevant definitions from the Microsoft PUR document below.

“Primary User” means the user who uses a Windows Software Assurance or Windows VDA Licensed Device more than 50% of the time in any 90 day period.

“Companion Device” means any additional device that is used by the Primary User, and either (i) is not capable of Running an Instance of Windows 8 Pro locally (in a Physical or Virtual OSE), or (ii) is both capable of Running an Instance of Windows 8 Pro locally (in a Physical or Virtual OSE) and personally owned by the Primary User.

“Window RT Companion Device” means a Companion Device you (not a third party) have licensed for Windows RT.


@Martin, what you say only applies to employee-owned devices. Corporate-purchased devices running Windows RT (such as Surface RT tablets) are entitled access to VDI environments without the need for an additional license. This is the topic that got all of this started.

Certainly end users wouldn't buy thin clients, but given the opportunity to avoid buying VDA licenses, a corporation would consider a situation like this.


@Gabe, no what I say doesn't just apply to employee-owned devices - the PUR document is the primary authoritative reference for how a corporation can use its licenses.

What I am saying is that document contains clauses allowing Windows RT devices to access VDI without additional licenses only if it is a "companion device" and that only applies if the user also has a device with SA or VDA.

If you can point to where it says that you can access VDI without being a companion device then that would be very interesting. If true then it could even make sense to use a Surface RT as a thin client just for the license savings.


Yep, you're right. I forgot about that little bit. Think of this article as a thought exercise :)

I'll make an edit to the article about this. Thanks so much for calling that out.