Desktop Linux has floundered for three main reasons: too few applications, limited desktop hardware compatibility, and too few tools (not to mention skilled people) to manage a boatload of Linux desktop systems.
Desktop virtualization, on the other hand, has had trouble getting off the ground because the value proposition falls just short of compelling. Sure, over the long haul, the benefits of centralized administration outweigh the investments you must make in server hardware, network bandwidth, and the transition to thin clients. But when you throw in the cost of VMware plus one Windows and Microsoft Office license per user, you may have trouble finding the "R" in "ROI."
[ Also on InfoWorld: Get a quick overview of one of the hottest trends in IT with "What desktop virtualization really means." | For a more detailed exploration, read Paul Venezia's "Thin Client Computing Deep Dive." ]
So what happens if you put desktop Linux and desktop virtualization together? If it were an all-open-source solution, the cost equation wouldn't even be in the same ballpark.
As InfoWorld executive editor Galen Gruman wrote last year, desktop Linux -- in particular, Ubuntu -- has come a long way in usability. And last April's release of Ubuntu 10.04 further improved the look and feel with the new, more polished Light interface. As for applications, there's OpenOffice, of course, but also version 3 of Lotus Symphony, which is IBM's most serious attempt yet at a free-as-in-beer alternative to Microsoft Office.
As for open source virtualization, to take one example, Red Hat updated its Enterprise Virtualization offering to include support for desktop virtualization last March. I hasten to add that we have not tested this package or KVM (kernel-based virtual machine), the open source hypervisor on which it is based, though the InfoWorld Test Center plans to do so this summer. I'm not making endorsements. What I'm just saying is that all the pieces are available for you to assemble.
What might happen when you put those pieces together? Well, in addition to vastly lower cost than a Microsoft-based solution, if you use thin clients, then very obviously all potential Linux compatibility problems with desktop hardware evaporate. You also pretty much eliminate all endpoint security vulnerabilities. I'm not naïve about the time and effort that might be required to get such a bleeding-edge VDI solution up and running, nor do I want to imagine what desktop video and audio problems might ensue. But I bet some sufficiently zealous Linux admins could get it all to work.