ACE, VDI and Other Dumb Ideas

analysis
Mar 6, 20072 mins

Yesterday's announcement by VMware that version 2.0 of their ACE virtual machine management platform had entered beta accomplished several goals. First, it helped VMware to upgrade the buzz factor surrounding its VDI initiative. It also helped to divert attention away from the company's recent ill-conceived swipe at Microsoft's licensing terms (see my recent post about this). But the primary goal - to cancel-out

Yesterday’s announcement by VMware that version 2.0 of their ACE virtual machine management platform had entered beta accomplished several goals. First, it helped VMware to upgrade the buzz factor surrounding its VDI initiative. It also helped to divert attention away from the company’s recent ill-conceived swipe at Microsoft’s licensing terms (see my recent post about this). But the primary goal – to cancel-out growing unease about the efficacy of the VDI model – remains elusive to them.

The truth is that IT shops aren’t buying into the whole “VM as a desktop replacement” story. These customers were burned once already by the server-based computing fiasco, so they’re understandably skittish about any solution that purports to replace the desktop with “something better.” The “something,” as they’ve discovered, is rarely “better,” and ACE/VDI is no exception. The idea that you can solve today’s desktop management problems by taking a “fat” client environment and isolating portions into a bunch of even “fatter” virtual machines – each with its own set of hardware incompatibilities and performance limitations – smells of bad science.

In fact, this entire scenario reminds me of a debate I had nearly 10 years ago with my analyst colleagues at Giga Information Group. Back then it was Terminal Services and Windows CE – a potent Kool-Aide that had nearly the entire analyst staff predicting the demise of the traditional desktop. Ever the contrarian, I countered with a vision centered around a Windows NT-derivative on the desktop and policy-based distribution of traditional “fat client” applications – my “Windows as a Platform” thesis (guess who was right).

The names may have changed, but the game remains the same. So this time around I’ll counter VMware’s sweet elixir with a dose of hard liquor reality: Think Windows Vista-derivative running streaming, virtualized application processes under a subscriber license model. It’s a not-so-bold, down-to-earth prediction, one that has the advantage of direct synergy with the stated goals of the ultimate authority on all things desktop: Microsoft.