I wrote on Sun's VDI upgrade earlier this month. Charlie Babcock interviewed Jeff Harvey from Cincinnati Bell about that company's VDI deployment...

While I see Sun's VDI as a pretty slick any-to-many broker for desktop virtualization, we all know Sun also is hoping to move a bunch of thin clients out the door. According to Harvey, CB's project manager, he'll be moving 750 Win2K desktops to Sun Rays. Not surprisingly, the target desks are mainly folks wearing headsets. Check out Charlie's InformationWeek piece here.

We've all heard this pitch before. Heck, I inherited a MetaFrame (pre-Presentation server, pre-XenApp) ICA cluster a couple of jobs back and had mixed personal feelings.

I can say:
1. My budgetary 'feelings' weren't mixed.
2. My admins and support staff loved the environment.
3. My users, um, didn't.

Have user attitudes changed over the last four years?

Do enterprises care about user feelings?

The cluster's operational costs were pretty much flat for five years. Those 30 seats are probably still rolling on the original terminals. We did one server SW + HW rev on my watch with no user downtime to "upgrade" all seats. Pointing all terminals to the new back end was seamless; users came to work and found a "faster" machine on the desk.

Sun's VDI looks to deliver a pretty robust user experience on a variety of user hardware: PCs, Macs, or terminals. One of the big benefits of VDI is the flexibility of its Java-based client app; sites can roll out a portfolio of desktops as appropriate to pretty much any flavor of hardware. Lock 'em down, leave 'em be, or give 'em an inexpensive screen, keyboard, and mouse. Let 'em work from home. It doesn't matter.

Sun Rays list from $249 up to $849 for an all-in-one with 17-inch LCD. VDI is licensed at $149 per concurrent seat. You'll need to contemplate your navel for a bit to see if those numbers make short- and long-term sense for your shop. It seems they do for Cincinnati Bell.

There has been a bunch of reader commentary to Charlie's piece, mostly on the relative ROI/TCO/etc. merits of terminal-based solutions. It's worth reading; well-defended viewpoints, strong opinions, some minor name calling, and solid peer feedback.

Whatever your opinion on the thin-client debate, we're all going to be seeing different flavors of desktop virtualization and respective delivery mechanisms becoming a (the?) major operating model in the not-so-distant future.