I wrote a blog article back in 2017—Windows 10 and Windows Server 2016 – the Old Battle of VDI Versus RDSH—wherein I discussed in depth the relationship between Windows 10 as a client OS and Windows Server with the Remote Desktop Session Host role installed.

Since virtual desktops became a thing, this debate has continued pretty much in the form that I laid out there. But since the start of 2019, there has been a slight sea-change in the environment, brought about by Microsoft’s Windows Virtual Desktop offering and the new iteration of Windows 10 that exists within it.

Let’s go into that now, and find out the answer to the very important question: Has the game finally changed?

The previous article referenced the differences between Windows Server (2016 or 2019) and the Windows 10 desktop OS, and how these translate into choices that you need to make for your own virtual desktop solutions. However, the advent of Microsoft’s Windows Virtual Desktop (WVD) offering has brought a new aspect to this old discussion: Windows 10 Enterprise Multi-User.

Windows 10 Enterprise Multi-User

Windows 10 Enterprise Multi-User is simply exactly what it says on the tin: a version of Windows 10 Enterprise with multi-user extensions. It exists because there has long been a feeling that Microsoft eventually want to retire the Remote Desktop Session Host role from their server operating systems. Microsoft have publicly stated: “We do not believe servers need, or should have, the personalized experiences that Windows 10 provides so well.”

It is true that Windows Server with RDSH is almost an anomaly that grew out of Microsoft’s on/off relationship with Citrix. Citrix provided the first ever multi-user components for Windows, and then Microsoft moved into the market themselves with NT4 Terminal Server edition.

But finally, Microsoft seem keen to evolve Windows Server as a real server-focused operating system, without the hangover of providing the capabilities of RDSH. Indeed, RDSH is one of only two server roles (I’m not entirely sure what the other one is, unfortunately) that still requires a GUI to be installed. In this containers-centric world of Docker and Kubernetes that we are moving into, it seems that Microsoft feel that having a specific role that turns their server product into a multi-user client operating system is not just archaic, but a headache in terms of functionality and development.

The RDSH role is still present in Windows Server 2019, and indeed, brings many optimizations for cloud deployments, but it seems very clear that it will not be a long-term goal. Server 2019 lacks some of Windows 10’s features, such as Cortana, the Microsoft Store, Edge, and originally, even Office 365 ProPlus support (a move Microsoft have now backed out of, albeit possibly temporarily).

Windows 10 Migration: Request A Demo

Windows 10: The Superior User Experience

Windows 10, even in the multi-user variant, provides all of this, because it is simply a slight variant of the same code base. It includes full current release updates, has all the features excluded from Server 2019, and Desktop Search is optimized for the user environment. Microsoft appear keen to position this as the best of both worldsthe scalability of RDSH, combined with the application compatibility, user experience and features of Windows 10.

This is possibly a very good thing. I’ve done many deployments where I have to provide a unified user experience across users both of Windows 10 and Server RDSH sessions, and the disparities between the two have often made these tasks unnecessarily complicated.

Say a user pins the Edge icon to their Windows 10 taskbar and then logs on to a Windows Server RDSH desktop. Do you put up with the fact that the Edge icon now shows as blank or missing, or do you maintain two sets of profiles, or do you try and work around it with a profile management solution that only restores the icon when you hit the Windows 10 OS?

What about new features? When users see and use a new feature on Windows 10, how do they manage when they realize the same feature is missing from RDSH? What if you have an application that runs on .Net Framework 1.1 – Windows 10 can cope, but Server can’t. Do you need UWP apps? There are a number of situations where marrying the multi-user and single-user operating systems together makes our jobs a darn sight more straightforward. Let’s not forget, an RDSH desktop or application isn’t a true client OS experience—it’s simply been made to look like one. And as I said earlier, the development of the Server OS as a vehicle to accommodate multiple users is a hangover from the late 90s, when VMs were a new thing and the extra raw compute power needed to handle multiple sessions was only addressable by server operating systems. Nowadays, just about everything is virtual, so why should we not look to bring client systems with multi-user capability back on the same level as those which are used in a dedicated, one-to-one fashion?

Microsoft Office Compatibility

But that’s not the only consideration - the Office compatibility is also a pretty big deal. Microsoft backtracked on their original decision to remove Office365 ProPlus compatibility from Server 2019—very possibly after facing a pushback from big enterprises with huge investments in RDSH deployments. You can now use Office perpetual or ProPlus on Server 2019. But the fact that this decision was made in the first place again leads you to catch a glimpse of Microsoft’s long-term strategy. Ideally, they want all client workloads (such as Office usage) moved away from the server OS and pushed towards Windows 10, whether it be multi-user or single-user. They may have relented for Server 2019, but as they ramp up the usage of Windows 10 multi-user as the long-term RDSH replacement, don’t be surprised to see them try to move in this direction again.

Windows 10 Multi-User Install vs. Traditional RDSH Install

So how does Windows 10 multi-user compare against a traditional RDSH install? Unfortunately there aren’t yet many clear comparison cases from the field, but the initial consensus is that Windows 10 multi-user seems to run, out-of-the-box, with slightly less users than you’d expect to see on an equivalent RDSH system. This is possibly to be expected, though, because not only does Windows Server have a slightly lower footprint than Windows 10, but we’ve also spent many years performing heavy optimization on RDSH desktops to extract every ounce of performance from them. Once the same optimizations are tuned and applied to Windows 10 multi-user, it may well be that we will get a more closely aligned level of performance.

But there is one big problem for this that we haven’t yet mentioned. Unfortunately, Windows 10 Enterprise Multi-User is only currently available on Azure as part of the Windows Virtual Desktop service. It’s actually against the license agreement to use it outside of Azure, and it won’t activate against on-premises KMS.

This is going to be, currently, the single piece that stops the game changing radically. If you’re not planning on investing into Windows Virtual Desktop, and you’re intending on keeping your multi-session hosts within a hosted or on-premises deployment, then you will need to adopt Windows Server 2019. This will get interesting for third-party DaaS providers, as they struggle to provide a true Windows 10 experience to match that provided by Microsoft. And if Microsoft finally choose to remove Office ProPlus support from later Server OS versions and essentially lock anyone who wants Windows 10 multi-user with the full version of Office squarely into an Azure-shaped coffin, then where does that leave us from a competitive standpoint?

I’m hoping that Microsoft don’t go down this route, and that they make the multi-session version of Windows 10 available beyond the Azure perimeter. Whether this is extending it to other cloud providers, making an Azure Stack or a fully on-premises version available, I just hope they don’t go for full lock-in. But given that Microsoft have finally moved their tanks onto the lawns of Citrix, VMware, Parallels and others, the developments around Windows 10 multi-user definitely point to an interesting few years ahead.


So in summary, will Windows 10 multi-user supersede and kill off RDSH? It certainly has the potential to, and aligning the multi-user version of Windows with the client OS rather than the server is a good idea for all concerned. But until Microsoft make it available beyond the shores of Windows Virtual Desktop, RDSH looks like it will grimly hang on. For now.

About the Author

James Rankin has over twenty years of experience in the IT industry, fifteen of which have been spent in the End User Computing area (EUC). He has a broad spectrum of skills centered around application, user, desktop and presentation virtualization. Check out his blog and follow him on Twitter.