It's clear that the industry's capability to deliver desktops hosted in the public cloud is maturing quickly, and while there's no chance anyone will call 2017 The Year of DaaS, we're starting to get a clearer picture about how desktop virtualization will evolve over the next decade or so. We're quickly getting to the point where the technology, cloud strategy, and mindsets are all good enough to move our desktops to the cloud, but the one thing we're missing is motivation. What will be the catalyst that drives our desktops out of our datacenter?
Until recently, I'd always believed that the desktops would be the last thing to move, though other people I've talked to have suggested that desktops will move first and drag the other datacenter-based services with them. (To be honest, one of those people works for a company that sells a cloud-based desktop virtualization product, so there may be a bit of marketing in there). That led me to start thinking more about the order of events, though, and I think I've reached a moment of clarity that allows both situations to evolve simultaneously.
Desktops vs Apps
Too often, RDSH-based applications and VDI are lumped together at high levels, usually under the term desktop, but to understand what's going to happen we need to separate the two. For the sake of this conversation, let's establish some definitions:
- "Desktops" are Windows desktops that we deliver to our end users. They have Start Menus, task bars, apps, icons, kitty mouse cursors, and pictures of grandkids or cars or supermodels for a background.
- "Applications" are published applications that we singularly deliver to our users for some strategic purpose.
This is important, because we already deal with each of these independently, combining them in the end to create our end users' workspaces. Separating them allows us to address the evolution of each one independently, and that's where things start to make sense.
Applications will go to the cloud first
There's a pretty good chance that you're already leveraging some cloud platform today. Maybe it's only for Office 365, but even then, you're probably considering putting other workloads in Azure. There's also a pretty good chance that desktops are not one of those workloads because you have so many resources in the local datacenter that it doesn't make sense to move the desktops to a geographically different location.
However, as you move certain datacenter-based workloads to the cloud, there will come a time when the applications you use to access those workloads will suffer from the same geography puzzle. If the application's data is in the cloud, wouldn't it make sense to put the application in the cloud, too? The answer to that is most likely, "Yes!" but not if it means moving the entire desktop and leaving the rest of your datacenter-based workloads in the datacenter.
This is why it's important to separate Applications from Desktops. In the scenario above, it makes no sense whatsoever to move your entire desktop environment to the cloud just so a single app can be close to its data. It does, however, make sense to deliver published applications from the same location as your cloud-based application data workload.
This is Phase 1 of Desktop Virtualization's move to the cloud. It's happening right now. As we pick a workload that has a user-facing application to move to the cloud, the application should move along with it. The application will be delivered from the cloud to a desktop running on-premises. Depending on the number of datacenter-based applications you have, you might be able to sustain this practice for a very long time.
Desktops will follow…eventually
Someday, probably way down the road, there will come a time when enough of your on-premises workloads are in the cloud that it finally makes sense to move your desktops there, too. Each company is different, so I'm not going to make a sweeping declaration that says everyone will have their desktops in the cloud by 2021 or anything. There is, however, a tipping point, and when you hit that point, you'll know.
What remains to be seen, though, is whether we'll care about the Windows desktops when the time comes to move them. Think about it, if you've slowly, successfully moved all your applications to cloud-based published applications, presumably you've solved a few other technical challenges along the way. You're no longer as dependent on the Windows desktop as you once were, so do you even need it around to begin with? Is the world even still using Windows desktops?
For the record, I think the world will still be using Windows because Microsoft has done a good job of keeping it relevant over the last year or two. iOS and Android aren't made for a desktop form factor and maybe never will be, and though ChromeOS, macOS, and Linux keep stealing fringe enterprise use cases here and there, it'll be a long time until they unseat the king.
So, when will your desktops move to the cloud? When you're ready! Until then, focus on applications, and look for platforms that make it easy to put published applications next to the cloud-based data they depend on.
Right now, all the platforms seem to be of the do-it-all variety, but I wouldn't be surprised to see providers that want to focus on desktops targeting SMEs that don't have a long list of applications that require resources in their datacenter. On the other hand, providers that want to take a chunk out of the large-scale enterprise market should focus on offering turnkey (or at least easy to build) solutions that enterprises can leverage as they move application workloads to the cloud.