Scott McNealy of Sun Microsystems has been talking about the concept of the "big freaking Webtone switch" for years now. (Check out this 2001 interview.) Essentially, he's been convinced -- and apparently still is, as Dan Farber's report from OracleWorld relates -- that Sun's future lies in the delivery of applications and processing power on an on-demand basis from the network (off of Sun servers, of course). This is also the model that drives IBM's vision. Sun and IBM are building it, but, as Dan asks, why aren't they coming?
I remember when IBM, Sun and Oracle ganged up on Microsoft in the late 1990s to push the thin client concept. It was a great idea, with a lot of great technology. But hardware prices took a plunge, bringing the prices of fully loaded end-user workstations down into the hundreds of dollars.
I often think that an analogy can be made between private automobile ownership versus public transportation. No matter how cheap, clean and efficient public transportation can be, a lot of people will still drive their cars. On some level, we prefer to pay a little extra for flexibility, and the ability to be masters of our own destinies (or more appropriately, destinations). Also, it's physically impossible for public transit systems to serve every path to work, every schedule, and every contingency.
Don't get me wrong; I'm a great believer in public transportation, as I am in subscription-based IT services. But as long as there is relatively cheap hardware, reasonably priced software (especially now with open source), and talented people around, most businesses will find it easier to get from point A to point B their own way.