BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Don Jones and Greg Shields, of Concentrated Technology firmly believe that business comes first, not technology. While they feel that VDI technologies are undeniably cool, they don't necessarily believe that they offer true value to a lot of the businesses that pursue them. In this series of articles, the Skeptics outline their arguments for and against various VDI technologies, helping you focus on what your business actually needs -- not necessarily on what a vendor or two might want to sell.
During every custom training class we teach on virtualization, someone in the group asks the question:
"So, uh, we're thinking about doing VDI. What are your thoughts?"
We hate that question. It forces us to pull out the full disclaimer that we are, after all, the VDI Skeptics. That label should immediately taint any of our opinions forthwith. But, with the question asked, our answer is always the same:
"We believe that VDI is not the answer. It's an answer, but it isn't the answer for everything the vendors would have you believe. The answer is, 'What do your applications require?'"
That vague response is purposely constructed to drive a further conversation, one that intends to re-frame the entire story line surrounding the reasons for VDI. You see, it's our opinion that VDI as the answer to desktop management is in fact an answer to the wrong question. VDI's intricacies, dependencies, hardware, and endpoint requirements are too-often missing calculations in a poorly-designed ROI. VDI, in our experience, doesn't generate an acceptable return when "getting rid of physical desktops" is the client's primary goal.
Being a VDI Skeptic doesn't mean we're anti-VDI -- far from it. VDI isn't without its use cases. Yet those use cases are often missed when the right questions don't get asked. Those questions relate to application delivery.
Framing 'application delivery'
Think for a minute about the applications your users work with every day. But don't focus on the application itself. Rather, consider the ways in which those applications are made available to users. Are they installed locally, either manually or through some automated software installation solution? Are they streamed to the desktop via application virtualization? Are they installed to a server and presented to the user via some remote protocol? Or, in the most heavyweight of options, are they installed or streamed to a virtualized desktop with users somehow connecting to that desktop or its application on top?
These questions highlight the four major application delivery approaches today's desktop virtualization technology supports. They also describe a spectrum of application delivery options that range from the operationally lightweight (direct installation) to the heavyweight (VDI).
Greg recently wrote that the core mission of IT is to create, manage and ensure secure access to business applications and data. Everything else, whether it be security or troubleshooting, help desk support or server administration, are all activities in support of that primary mission. At the end of the day, IT's reason for being is to ensure end users can securely access their applications and data. How we manifest that access is a task left up to us.
Going a step further, if secure access to applications and data is the primary charter of IT, then the aforementioned "correct" questions must all start with the applications themselves. Think about what those correct questions might be: Can this application be installed or streamed to local hardware? Is remote access required? If so, can I install it to a server and present it to users through a session? If the app doesn't work in a session environment, what do I gain by installing it to a virtual desktop and presenting that desktop to my users?
We all know that in an apples-to-apples comparison a Remote Desktop Services or XenApp Server can support a far greater user density than any equivalent VDI host. That's just the nature of sessions. But, some applications don't work well atop RDS, XenApp, or any of the other session-based application delivery mechanisms available today. Whether their reason relates to performance, poor coding, or a bad installation routine, it is the problem applications that make VDI a potentially smart move.
Seeking the use case
Admittedly, there are other VDI use cases that aren't completely application-dependent: The computer lab that requires a complete re-build with only ten minutes between classes, the power user who requires a personalized desktop when they are on the road, and a myriad of others. All are specialized situations where the need for full desktops trumps the incremental cost required to get them. For everything else, lighter-weight options exist that bring the same security and centralization VDI extols, but at a lower cost of ownership.
If you still disbelieve the primacy of this whole notion of application delivery, consider how the operating system itself has become little more than a container for applications and data. Once the central focus of IT, today's OS has arguably become the least important part of the entire IT infrastructure. It has, in a way, become the digital equivalent of a building's electrical and water systems. You absolutely need them, but you don't think about them most of the time, at least until something breaks.
The rapid embrace of new computing form factors further highlights the demise of the traditional OS role. Tablet and mobile computing across a wide array of form factors and operating systems presents a growing set of targets for the delivery of applications. With the right hardware, that delivery can finally be right-sized to the needs of its user and the tasks they're accomplishing.
VDI isn't the answer to those users' needs. It's indeed an answer, but only one in a range of opportunities. No, the future of IT is all about its applications. Being able to deliver those applications in whatever format users need is the key skill IT is only now beginning to figure out.