Recently I was asked, “Don’t you think it’s weird that Microsoft is talking all about Azure on one hand, and then pushing a desktop OS so hard on the other?” At first blush, it might seem strange, but we’re not talking about a new startup trying to do too much, we’re talking about Microsoft. Their push towards the cloud is exactly what they need for future relevance, but they can’t ignore Windows and what it’s done for (and to) them. At various times throughout its history, the world has either gravitated towards or away from Microsoft based on Windows. Basically, to get people to the cloud, Windows also has to be successful.
Do you think companies that rely on Windows (nearly all of them) would be willing to move core services and applications to Microsoft’s cloud in the face of stiff competition if they let Windows and its apps languish in the corner while focusing all their efforts on on the sexy new cloud? Not at all. The phrase “Dance with the one brought you” comes to mind. Microsoft realizes that the way to bring people into Azure is by showing the world that they have their collective act together. Then they can hold customers’ hands as they catch up. It’s a strategy shift: Don’t give them the OS that you want them to have (Windows 8), give them the OS that they want to have (Windows 10) before it’s too late and you lose them forever.
But that’s not all. Windows isn’t just the first taste that gets people coming back for more. Microsoft knows as well as we do that the reason we keep Windows around is for the applications, and that we will be managing Windows apps (and because of that, Windows) for the foreseeable future. The ongoing dependence on a desktop OS to run these applications can only partially be addressed with Azure RemoteApp, so I think we’re going to see Microsoft do a harder push towards a real Azure-based DaaS platform in the near future.
Some might think it’s strange that they haven’t done it yet, but when is Microsoft ever on the forefront of technology? Even the stuff that we use in the desktop virtualization space is all stuff that was acquired and/or adapted. RemoteFX came from Calista. RDS evolved from the work Citrix did back in the 90’s with MetaFrame and Windows NT Terminal Server Edition. Hyper-V was mostly net-new technology from Microsoft (though some came from Connectix and Virtual Server), but it was released well after VMware had established their lead.
My point is that Microsoft, at least in our desktop world, doesn’t lead; they watch, they tweak, and then they execute…well. They’re going to let all the other providers learn the hard lessons while they do some housekeeping and get ready to make an impact. The last two years have been spent watching and tweaking. First, we got Azure RemoteApp, which instantly sussed out potential customers that already had the kind of mindset needed to move to DaaS. The focus on applications meant that Microsoft didn’t have to sell customers on moving to Windows 8 as part of the deal since it was quite clear by then that Windows 8 was a failure.
Then they created a per-user VDA license, which doesn’t really help service providers but does open the door to a subscription-based, per-user license should Microsoft decide to go down that path. To truly embrace service providers, they’d have to allow multi-tenancy, too, but we’re in the tweaking phase right now, not the execution phase (more on that later).
While this was going on, Microsoft watched along with the rest of us as service providers struggled to sell big customers on the new delivery model. DaaS is a tough pill to swallow (for reasons we outlined in the book), and where the providers that jumped on early had to come up with offerings that they thought were best, the rest of us could watch the reactions of the industry and plan around it. The overwhelming reaction: We like the management, flexibility, and incremental scaling, but we want our stuff on-premises or we’ll just keep doing things the old way.
Enter Azure Stack, which was announced at Microsoft Ignite in May 2015. There’s no mention of DaaS in all the talk around Azure Stack, but the description from the introductory TechNet blog post sounds exactly like what potential customers say they need:
Azure Stack delivers IaaS and PaaS services into your datacenter so you can easily blend your enterprise applications such as SQL Server, SharePoint, and Exchange with modern distributed applications and services while maintaining centralized oversight.
Meanwhile, the Windows team was heads down, working on Windows 10 and getting ready for the launch that just occurred a few weeks ago. Though the launch wasn’t met with much fanfare, I’ve yet to hear a single bad thing about Windows 10. Most people are cautiously optimistic, and just about everyone has some sort of a plan for bringing it into their organization. Some are rolling it out as they get new hardware, others are waiting for a certain LTSB build (the days of waiting for the first Service Pack are over!), but seemingly everyone has a plan. Once widespread adoption of Windows 10 is confirmed and Microsoft no longer has to “sell” any customers on the OS as well as the delivery mechanism, they can move on to the execute phase of their DaaS plan.
All that’s really needed to launch a DaaS platform now is a licensing change. In order to play by their own rules today, a Microsoft DaaS platform would have to have separate compute hardware for each customer. Surely they realize how asinine that is (though, per El Reg, maybe they don't), but it hasn’t really affected them compared to how it’s held back the service provider ecosystem. In fact, I could probably make a case for Microsoft intentionally leaving licensing as it is until they were ready to enter the fray as well. It sounds villainous, but it’s well within their rights. I think we’ll see Microsoft making a big licensing change at or at about the same time as they release a Windows 10-based DaaS platform. (Frankly, the combination of Azure Stack with Windows 10 could get around SPLA licensing since all the compute resources would be dedicated to a single customer, but that’s a limited use case.)
Just when you thought DaaS was dead, it appears that Microsoft could be poised to make a go of it. When Microsoft tries to drive innovation, as they did with Windows 8, they haven’t been all that successful, but give them time and resources and they can execute with the best of them. Will we see a DaaS platform from Microsoft in the next 6-12 months? I think so. What about you?
How to license Windows 10 virtual desktops