Citrix, VMware, and other vendors are spending millions trying to convince customers that they need VDI. The reality is that everyone who really needs VDI is already using it.
There are 500-700m corporate desktops in the world. The number of VDI users is not known for sure, although it's thought to be in the single-digit millions. (For sure less than 5m, probably less than 2m.) So why aren't the other 498m users in the world using VDI? Simple. Because they don't need it!
Look, if they needed it, they'd have done it by now. Look at what Harry Labana did at Goldman Sachs when he built his own VDI out of single-user Citrix Presentation Servers running as VMs. Look at the financial companies using VDI. Look at the people with outsourced developers.
If you need it, you're already doing it. If you're not already doing it, then you don't need it!
So now what? I guess we move on to those who want VDI. Or to those who think it might be nice.
But there's nothing standing in the way of those who need VDI from using it now. I mean what are you waiting for? The protocols aren't getting any better. The management isn't getting any better. You don't need the single disk sharing to work perfectly (since you can use TS for that). So if you need VDI, get on it!
While I wish I could've just ended this post at that, I should point out that just because those who really need VDI are already using it, this doesn't mean that the vendors have nothing more to sell. Client-based VMs are going to be huge in 2010, as will the various data replication and syncing tools to ensure that users can run VMs locally as easy as they can access them remotely. These technologies will usher in a new wave of adopters (again, those who need them), which will in turn push a whole new set of technologies. The cycle will repeat.
It's just that right now we're sort of stuck in the "VDI-is-as-good-as-it's-going-to-get-phase-and-client-hypervisors-won't-be-out-for-another-six-months" doldrums.