Businesses like cutting costs, and they will spend money on things that can help them achieve that. This goes a long way toward explaining why so many organizations are investing in server virtualization. However, conspicuous by its absence is any trend toward organizations investing in desktop virtualization technology, or VDI.
Why should that be? One answer is that it doesn't help cut costs. A less obvious one is that most organizations don't really need it.
It turns out that you're pretty much out of luck if what you want from VDI is cost savings. Take a look at the websites of some of the key VDI products--Citrix XenDesktop, Microsoft RemoteFX, Quest vWorkspace and VMware View--and you'll see plenty of benefits from VDI, but not much in the way of cost savings.
According to Citrix, Microsoft, VMware and Quest, the benefits you'll get are:
- Increased workforce productivity
- Hardware independence
- The ability to use the latest mobile devices
- Business continuity
- Organizational agility
- Improved manageability
- Increased reliability
- Fast, flexible desktop and app delivery for branch expansion and other initiatives
- Enhanced security
Not much in the way of cost-cutting capabilities, is there? In fact it's only VMware that makes any obvious claims to savings: "Lower your IT operational expenses by up to 60 percent," the company teases. That's a big claim, or it would be if it weren't for the "up to" bit.
But does VDI really save you money? Not normally, according to Andy Paul, principal virtualization consultant for GlassHouse Technologies, a Massachusetts-based data center infrastructure consulting company. He reckons that in most cases companies are being misled if they think that they are going to get a direct financial boost from implementing VDI. "Except in the biggest implementations, VDI is not a cost-saving exercise," he said.
One of the reasons for this is the heavy upfront costs of any implementation. This includes the virtualization host hardware, software licenses, storage resources, upgraded networking capabilities and data center fees.
Of course, some of these costs are offset by extending the life of obsolete client hardware and the fact that there is less hardware to buy, manage and maintain in office environments. "When VDI is done correctly, VDI probably ends up being net neutral as far as costs are concerned--your increased up-front capex costs are offset by reduced day-to-day costs," said Paul. But in most cases it will take something like five years before a VDI project pays for itself, he believes, by which time it will probably be time for a new project.
Still, that's not to say VDI is without merit--it does offer other benefits, although the value of some of these are hard to quantify. One of the most obvious is security: The security from having all your data safely stored in a data center where it is unlikely to be lost or stolen--unlike the case when it is stored on a laptop, or even a desktop in an unattended office. And there are other benefits too--things that Paul calls management-layer items, like easier upgrading and patch management.
For the virtualization technology vendors like VMware, Microsoft, Quest and Citrix that are pushing VDI solutions, the lack of a solid ROI story is bad for sales. But there's another problem which must also be gnawing at them: Is the desktop form of virtualization really necessary in very many cases?
There are alternatives, after all. In many situations it may be that users need access to a single application, so an app virtualization solution like Citrix XenApp or Microsoft App-V may be more appropriate than full desktop virtualization. And far less costly in terms of upfront expenditure. Then there's also good old-fashioned server-based computing, which is vastly more mature while offering many of the same benefits, including better security. "All things being equal I would steer people to server-based computing because you can achieve much better density," said Paul. "Server side computing is a form of VDI that is being overlooked at the moment, but in many scenarios it is up to the task."
So if you're considering desktop virtualization technologies, consider them with an open mind. Be aware that you might not achieve any cost savings you may be promised, and don't forget that whatever benefits you're after, desktop virtualization may not be the only solution that can provide them.
Paul Rubens is a journalist based in Marlow on Thames, England. He has been programming, tinkering and generally sitting in front of computer screens since his first encounter with a DEC PDP-11 in 1979.