Yet the fate of virtual desktops seems less assured than the vendors would have it, given casual conversations I had with attendees at VMworld a few weeks ago. Most of the people were there to learn, and wanted to be “more virtualized,” as if 100% virtualization was a laudatory goal. But what I took away from sessions and discussions was that businesses should start the VDI conversion slowly and thoughtfully, with non-mission-critical apps first. The oft-repeated disclaimer was that VDI may not work for every application. The downside is disconnection and latency, which renders employees less productive; and that costs a whole lot more than the VDI hardware.

Other reality checks are coming in. “If time is money, then, in my anecdotal view, this is a huge money hole,” writes a senior programmer analyst in response to my story last week on the ROI of VDI. “I do not have quantitative numbers to give you, but I would guess I am 100 times more productive on my old laptop than on the VDI environment. . . . I am excluding the number of times the VDI is down, or my session is unexpectedly terminated.” The performance is significantly slower, he adds. “Any action or movement by your mouse, or by entering in keystrokes, adds 5 seconds. . . . [A] problem that used to take 15 minutes to resolve will now take about an hour because I have to wait for the desktop to respond.”

While the virtualization industry works to improve such performance issues, significant growth in desktop virtualization has not been realized, according to IDC. “Vendors would need to continuously improve and simplify the [virtual desktop infrastructure] solution, and customers would need to understand client virtualization technologies and how to extract value from each component,” IDC concluded in a recent report predicting that client virtualization will begin to experience rapid adoption in the latter part of 2010 and in 2011.

Email me at lsmith@techtarget.com.