USAToday has a piece out today talking about the potential value and problems of virtual desktop infrastructures.
I know. I was shocked, too.
That's just my journalism snobbery talking, though. The article itself isn't aimed at the IT cognoscenti, but it's not written in crayon, either.
Despite surveys showing huge growth potential and projecting a very fast rollout, desktop virtualization in its many forms has been a disappointment -- at least for vendors hoping to sell more of it -- for at least the last four years.
Burton Group/Gartner virtualization guru Chris Wolf says the hesitation on the part of IT execs is that virtualizing PCs is a complex, long-term commitment, so they're making their product decisions carefully and doing their implementations deliberately.
IDC's Ian Song said cost is the big difference, but not just the cost of VDI.
Comparing the huge return most companies got from virtualizing servers -- getting rid of a lot of hardware and consolidating the rest onto more easily managed, newer, more centralized machines -- makes the relatively thin direct payback of VDI less attractive.
Just picking a flavor of virtual desktop is a trick, but then you have to decide whether to use old PCs rather than new thin-client hardware, you still have to license the software, lay out the network of servers that will feed all the endpoints their data, apps and OSes, figure out who needs a full-bore virtual-desktop running on your server and who only needs access to a couple of server-based apps with Web GUIs, and whether you want to try to put hypervisors on laptops and deal with user confusion over where "their" stuff went when they launch the 'home' VM instead of the 'work VM' or don't bother to run the VM at all and go to town with potentially sensitive data stored outside the secure-but-inconvenient VM environment on their laptops.
Desktop virtualization is complicated, and not only in its technology or economics. You're making complex technical decisions that affect nearly every move employees make using a PC. For them, that's a very personal situation.
It doesn't matter how much money you save if you put everyone on VMs, and as a result, their apps slow to a crawl because a virtual client can't handle the graphics the way it should, end users can't save or retrieve their settings so they must spend the first half hour of every day changing their wallpaper, fonts and screensavers, or the network was designed for email rather than for apps with a lot tolerance for latency or because your capacity-management calculations didn't take into account end users launching two or three instances of their desktops, or taking more than one virtual desktop option so they can work on the streaming-apps version while waiting for the VDI to load and figuring out how to get back into the resident VM on their laptop hard drive.
The end result should be a big savings on the maintenance front and give end users a better experience because you'll upgrade the server more often than you would refresh every piece of endpoint hardware in the business.
The problem with VDI, unlike cloud, SAAS, server and application virtualization, is that the combination of changes in hardware, software, networks and end-user experience makes the whole equation so complicated it may never make sense for big companies to do more than update and expand user bases that have always used terminal services, and sprinkle in a few or a lot of streaming or SAAS apps for everyone else.
The end result won't be a company running entirely on virtualized desktops, but one in which every end user has a desktop with some virtualization to it.