The future of virtual desktops seems a little murky right now. In 2008 Gartner predicted that 40% of desktops would be virtualized by 2010. They continued to predict this target over the next several years and were still maintaining as late as 2012 that 40% of corporate desktops would be virtualized by 2013.
After several years of stretching for this goal, in 2012 the rubber band seemed to snap. Brian Madden, one of the top VDI pundits, published his book “The VDI Delusion”. The book seemed to claim that although desktop virtualization was a good idea, the corporate desktop had beaten the VDI experts. Corporate desktops were by and large just too complicated to be virtualized.
Complication leads to a lack of automation. Lack of standards leads to increases in hardware spending. In the end, many felt that virtual desktops were just too hard and too expensive for the real world.
In 2015 Gartner is now targeting 8% of desktops to be virtualized by 2017.
Although it might be fun to turn this into a Gartner bashing blog, I don’t think that’s especially useful. The Gartner process is simple. They poll the industry and report back whatever we’re telling them about the future. It’s a little like polling graduating high school seniors and then confidently reporting that in the future nearly everybody will either be a pop star, movie actor, or the CEO of a technology company.
For a number of reasons, the industry wanted VDI to work out and for most people it has not. I think it’s worthwhile to do a little post mortem here.
There were a number of factors driving VDI adoption. VDI desktops were expected to be cheaper, more compliant, and easier to manage. They were also expected to be a great way to handle the explosion of possible end points in a BYOD or mobile world.
They aren’t cheaper and this alone tolled the death knell for many VDI initiatives. If you were under-provisioned, the whole thing fell in on itself. They are more compliant, as the data tends to stay in the data center and applications need to be intentionally provisioned. If you’ve got a good application provisioning system in place, they aren’t hard to manage. If you’re expecting to allow users to self-provision or you had to build application provisioning expertise just to support VDI, that probably caused problems. Death knell number 2.
In the end, VDI initiatives failed in good part based on faulty assumptions we had about how well-controlled the corporate desktop actually is. We found out that most organizations really don’t have a centralized way of provisioning all applications, just most applications. Most organizations really have no way to manage and move user personality settings. Many users have one computer in the whole wide world that they really like and are productive with. That computer has the exact applications they want, set up in the exact way that the user likes. Because the user has done it themselves. And that special, wonderful, highly personalized computing device sits comfortably on their corporate desktop.
But, with or without VDI, where do we stand? Our compliance and mobility needs aren’t getting any easier and we’re going to need some sort of solution.
Here’s the sad fact of the matter. It’s not just VDI that’s dead, it’s the whole concept of a corporate desktop.
Citrix has begun talking about the need for a service delivery fabric. A service delivery fabric recognizes the need to provision and control applications from a variety of originating points out to a variety of end points.
The origin of our corporate applications is no longer just the data center. It’s our data center. It’s IaaS. It’s cloud. It’s SaaS. End points are every possible device in very possible location. Sometimes they’ll look like traditional corporate desktops. Sometimes they’ll be mobile devices in a Starbucks in Sri Lanka.
The fabric is the control point for compliance and distribution. We need to secure the end point but we need to stop thinking of a corporate desktop standard as the control point for application provisioning. We need a service delivery fabric.
And, just to close the loop, is VDI going to be a part of that fabric? You bet it is. To handle the variety of origins and end points we’re dealing with, we’re best served with a number of tools. VDI will be one of them, just like SaaS, physical desktops, and iPads.
VDI as Gartner meme may be dead, but long live virtual desktops and their important role in the service delivery fabric.