In past blogs, I’ve discussed some of the big-picture benefits for server virtualization. This week, I wanted to give some thought to the concept of desktop virtualization. Of the two technologies, the desktop variety has been much slower to take off in the marketplace, and frankly, there are some good reasons for that.

In both cases, the fundamental underlying technology operates on the same basic principle: By uncoupling the operating system and application software layers from the binary code layer that actually operates the chip-level hardware, you can accomplish some cool things. Perhaps the number one benefit is being able to create the illusion of multiple independent physical machines (AKA “virtual” machines) all living side by side on a single hardware platform. These virtual machines can lead diverse and independent lives, being dedicated to various tasks and even running on different operating systems.

The benefits at the server level are almost self evident. For one thing, running multiple virtual machines on a single box can be more efficient, because you can eliminate underutilized machines. As a result, you can reduce hardware costs, eliminate redundancy, and slash you electric bill. Support and maintenance costs can also be reduced, as you consolidate your data center.

So what’s the value proposition for doing this at the desktop level? For me, at least, the picture is a little murkier. If you assume one user per desktop, which is the norm in most computing environments today, then the economies mentioned above for servers no longer apply. You can virtualize any number of desktop machines one a single machine, and give users access to those machines over the network, but you still have to put some kind of physical device on each user’s desk. These devices can be slimmed down, stripped of many of the features commonly included in a PC, but with hardware costs so low these days, you’d really have to sharpen your pencil to uncover significant and compelling savings in the hardware department.

And the savings must be compelling, because users will fight you every step of the way. Ever since IT came along and replaced our dumb terminals with PCs, we have been fighting tooth and nail to hold onto them. It’s human nature. We want the illusion of independence and choice, the ability to plug in our iPods and other peripherals, to be able to load our own apps and upgrades, and to generally make IT people miserable by continuously trying to corrupt their ideal world of a pristine, locked-down and homogenous desktop environment.

I’m offering a very simplistic view here, and I’m counting on knowledgeable readers to come along and set me straight with their cogent comments. In reality, there are tons of good reasons to virtualize desktops, and in fact, the technology to do so has been around quite a while and only gets better with age. But the implementations will continue to be limited, and most users will only go along grudgingly, until vendors and IT managers start selling the benefits directly to those end users, and can actually succeed in creating grass roots demand for virtualizing their desktops.

So how do you get users to the point where they are actually calling their IT managers and begging to be virtualized? Well, here’s what sold me on the idea, personally. I was rereading an excellent 2007 post from my old friend and colleague David Berlind, where he discussed his own experiments in desktop virtualization. His driving motivation was to be able to take his personal desktop and make it truly portable and crash-proof by being able to replicate it on pretty much any piece of hardware, anytime and anywhere. For me, at least, the light bulb finally went on.

Anybody who has had their hard drive crash and had to start from scratch knows the pain involved. Even someone who religiously backs up their data knows that unless you migrate to another computer with the exact same hardware configuration, recreating your comfortable old desktop environment is a laborious, painstaking and imperfect process. The ugly truth is, you can never really go back. There are some benefits of getting a fresh start every few years, but these transitions seldom come at opportune and convenient times.

But what if your entire personal computing environment – including even the subtle little things like software and security settings in your browser and related apps, and the look-ahead cache in your e-mail client that knows the names of the people you routinely correspond with and fills it in for you –all that was being stored as a complete software image that could be easily transplanted onto another machine at a moment’s notice without the slightest interruption or flicker of change or incompatibility? Now you’re talking!

I believe this is the kind of benefit that would truly resonate with end users, and start creating pull-through demand for virtualization at the desktop level. I truthfully can’t foresee a time in the very near future when the desktop virtualization market will outstrip its counterpart on the server side, but it can begin to happen once the dialog about why virtualization works finally moves beyond the server room and reaches the water cooler crowd. Of course, the software will have to be completely seamless and invisible to the end user, which includes paying no discernible performance premium for introducing that layer of abstraction between the user’s virtual machine and the actual hardware it’s running on.

So tell me, what am I missing?