As the saying goes, “Two things are certain in life. We pay taxes, and we die.” One of my more nerve striking posts and point of views is on the death of virtual desktop infrastructure VDI. I understand the tension. People have banked their careers and in some cases their entire business model on VDI. To say that their investment and sacrifice is just a short term investment is a very charged statement.
My LinkedIn Profile
I guess I should clarify my prediction a bit. VDI is dying; it’s not dead. VDI isn’t the long term vision of any major vendor for end user computing. Citrix is the king of VDI and has looked to diversify into server virtualization, cloud computing and other data center technologies such as load balancers. It’s ironic until recently one of my main personal use cases for VDI was VMware’s own vSphere Client on my Macbook. VMware has been pushing their vSphere Web Client as the default interface to manage their virtualization platform.
I think this points to the future of end user computing. Vendors will gradually web enable their applications or come up with completely different application models that make the local desktop OS irrelevant. But this is a long time away. As readers on my last post pointed out, there are still organizations relying on applications designed to run on Windows 2000 and IE 6. These organizations will need a bridge to their future computing platform, and this is where VDI fits in nicely.
However, there is a tipping point. There will be a point where most organizations will no longer rely on legacy desktop applications. There may be a need for some VDI, but the majority of end-user facing applications will be platform agnostic. Some people guess this will take 10 years or more. I don’t know. Technology has a tendency of going through evolutionary change quickly. There was a short period of time where I’d wire a desktop with both Ethernet and Token Ring to support direct connectivity to mainframe applications. The thought process was the mainframe would never go away. This was true the mainframe didn’t go away. But companies like Novell figured a way to proxy the connection to the mainframe and eliminate the expense of a second network.
VDI as it is today is a painful end user experience. An end user that purchases a Chromebook, iPad, or Macbook is not looking for a Windows experience. If they wanted a Windows experience, then the Surface would be a really hot selling device. My bet is that before this 10 year period is up, the industry will crack this nut. We’ll look at VDI as it is today and wonder why we ever thought that even the bridge period would have been presenting Windows running in a DC via remote control.
DaaS – A better option?
So where do we invest our time and resources if VDI isn’t the long term solution? As both an infrastructure professional and an end user, I get the appeal of VDI. It’s an “easy” solution to a difficult problem. It feels like the right solution to a very practical problem. But, this is where the engineer needs to meet the business and know how to advise the business in making the best investment. I believe short term that VDI is a necessary evil. I don’t see a practical way around it. However, if I were in a position of needing VDI, I’d seriously consider DaaS as a solution if it fits my use case.
Solutions like VMware’s Desktone can offer the best of both world’s. VMware takes the headache of managing the infrastructure while offering the advantages of an on-prim solution. I’m sure AWS isn’t too far away from offering an appealing solution as well. It helps control your exit (and you should have an exit plan for DaaS) from VDI. In theory, it gives the organization an elastic VDI bridge. When the industry figures out a VDI alternative you can ease back on the VDI infrastructure, reclaiming some of the opex associated with VDI. This should make the transition palatable from an economic and organizational perspective.
So, what are your thoughts? Is VDI that long term proxy or is it Token Ring?
Follow me on Twitter @virtualizedgeek