The perfect storm of bad news for VDI

High deployment costs, lack of multimedia support, pricey Microsoft licensing, and the Windows 7 transition all conspire to make the "obvious" desktop client alternative not so obvious

It began so innocently. While installing a new ESXi server and checking on a few racks of gear in a remote office, the office manager asked me politely if I could set up a workstation for the new intern. I haven't set up a workstation in many, many moons, but as "the IT guy," I was the obvious candidate. I smiled and said, "No problem."

After 45 minutes, two "perfectly good, just imaged" workstations, a bad patch cable, a Windows Active Directory computer object/SID overlap, and a slight headache from bumping my head on the desk, the XP workstation was finally up and on the network -- with the intern standing behind me the whole time. "Annoyed" is far too simple a term to accurately convey my mood.

[ InfoWorld's Randall C. Kennedy argues that VMware has lost its mojo. | VDI is missing five key attributes, as Paul Venezia outlines. | InfoWorld Test Center reviews: Citrix XenDesktop, Microsoft Hyper-V, and VMware VI3. ]

VDI would be perfect, except ...
Even more annoying and ironic, there was a perfectly good VDI pilot project running at the main datacenter. Well, "perfectly good," as in it's as good as VDI can get for the moment. Unfortunately, VDI is caught in a perfect storm of bad news.

One might think that the economy would be good for VDI. After all, it's supposed to save money, right? Well, yeah: over the course of a few years, following the initial purchase and implementation costs. But now really isn't the best time for many companies to be spending loads of dough on a new and relatively unproven desktop delivery mechanism. And yes, it will eventually save money, but you need to leverage what you currently have. Spending $300 or so on a thin client for every user will raise some eyebrows when you can get full-on desktop systems for the same money, without the expensive back-end infrastructure.

Also, VDI promises to save time and effort -- as long as your users don't need to use any form of multimedia or Flash-based applications. While many vendors have introduced relatively solid sideband video delivery tools to evade the nastiness that is video over RDP, these tools generally cost more, as do the thin clients that can use them.

And what of those existing desktops? Wouldn't it be nice to just turn them all into VDI clients? Well, you can by using the open source Thinstation or VMware's VDI Open Client, but unless you're comfortable heading in that direction, it's not a real option. And those tools do not have the multimedia redirection capabilities that the commercial thin client products offer.

But at least you can deliver a stable Windows XP desktop, right? Yes, for sure. Except that nobody really knows what's going on with Windows right now. Companies that moved to Vista are regretting that decision, and the rest of the world is waiting, hoping that Microsoft will finally get it right on its second desktop OS attempt in nearly a decade (a recently discovered bug in the forthcoming Windows 7 raises some doubts). Naturally, there are significant differences between XP and Windows Vista/Windows 7 in a VDI environment -- so much so that it's rather pointless to develop a pilot project on XP now because you'll have to redo the whole shebang for Windows 7, which will become available on Oct. 22.

Add Microsoft's draconian licensing for Windows XP VDI to the mixture, and it's basically a nonstarter for most infrastructures. The company may have changed things with Vista and Windows 7 licensing, but Redmond can switch it again at a whim. That will certainly make the compliance team smile.

If you're putting in a call center with only one or two fixed applications, then VDI might work out well, but for general-purpose office use, there are simply too many unresolved issues surrounding the technology to allow it to gain traction. It's almost as if certain players in the market are purposefully stunting VDI's growth until such time as they have their own VDI product to market. But I don't want to engage in conspiracy theories.

Time will undoubtedly heal these wounds (and presumably the bump on the back of my head), but it's anyone's guess how long that will take. Like it or not, the desktop dinosaur of the enterprise will remain with us for some time to come.

And we'll be free
Some day soon, and it's a-gonna be one day ...

Copyright © 2009 IDG Communications, Inc.

How to choose a low-code development platform