When we had Jon Rolls on the BrianMadden.com Podcast last month, our conversation turned (as most seem to do) to the future of Windows. In this case, it wasn’t about the future of Windows on the endpoint, rather, we talked about the future of Windows Server, and how it seems that Microsoft is increasingly focused on Nano Server as a huge part of their overall strategy.
Nano makes sense in the cloud, and you can make a strong case for it on-premises as well, due in no small part to the fact that we simply don’t require Windows to build applications for…well…Windows. There was a time when providing a GUI made a ton of sense, but for nearly all use cases these days, the one-GUI-per-VM (and all the other services that go along with it) is both unnecessary and crippling to you IT operations.
Without getting into all the details, mainly since they’re covered in so many other places (including here on BrianMadden.com), Nano is a virtualized infrastructure’s dream. Two years ago, Microsoft claimed it reduces VM size by 93%, has a 90% smaller security footprint, and requires 80% fewer reboots. With stats like that, why wouldn’t IT shift any application workload they could to Nano?
Though there are plenty of reasons IT departments aren’t in a hurry (32-bit workloads is a big one since Nano does not include WoW64 and many of those older applications rely on a local GUI instead of remote management), there is one broad exception that I can think of to this new age way of thinking: RDSH. To put it simply, without a full installation of Windows Server that includes an interactive session, there is no RDSH. How does that jive with what appears to the future of Microsoft’s datacenter operating system?
Maybe it doesn’t.
If Microsoft really is all-in on Nano as the future of Windows Server (which remains to be seen, but it certainly seems that way), what happens when the only reason they have left for keeping an interactive version of Windows Server around is essentially to keep alive a platform so that end users can access legacy applications? It seems silly to devote resources to develop and support the full Windows Server environment, RDS, and security updates when there is a perfectly good alternative that already does nearly all of that: Windows 10.
What if Microsoft decided to make Windows 10 multi-user? Keep in mind, I don’t think anyone is in a hurry to do this, but signs of this eventually happening are everywhere if you look hard enough. We’ve covered the emergence of Nano already, but if we look at the changes Microsoft is making towards licensing Windows 10, we can see an opportunity that did not exist years ago. Windows 10 already supports per-user licensing, so that’s taken care of. It now supports running in multi-tenant environments, too. That’s the same as SPLA, but we’re getting there.
It’s been rumored since Windows 2000 that the client version of Windows is “only a few DLL’s away” from being able to be used as an RDSH server. After all, Windows client OSes have been multi-user for years, and there are/were hacks that leverage Fast User Switching to enable an RDSH-like functionality. The point is, it’s possible that Microsoft would barely have to lift a finger to make Windows 10 the go-to platform for interactive Windows sessions. For Microsoft, that would mean they’d only have to support a single platform that featured an interactive session. For us, it would mean we’d also only have to support one interactive OS, which would simplify management and application testing.
I want to be clear that I’m not predicting anything drastic in the near future. I thought it would be an interesting discussion to start, though, so I’d love to hear your thoughts. Here are some more things to chew on:
- What would Citrix/VMware/others do? For example, would you install XenApp or Horizon Apps on Windows 10? It sounds weird, but there’s probably not any reason it can’t be done. Citrix may be better positioned to do it since they’ve already got RemotePC, but nothing is impossible.
- Microsoft has already scaled back the desktop experience on Server 2016 by removing Edge and Windows Store (though this is to keep Server 2016 compatible with Microsoft’s LTSB support model for Windows Server).
- In a way this is already happening today, we just use single instances of Windows 10 and call it VDI.
So what do you think? Does RDSH’s future involve Windows 10? Will Windows Server 2016 be the last to include a GUI? I can’t really shoot any holes in the idea at a high level. It’s possible that more fundamental issues could arise, but for the moment the whole idea seems to hold water.