Cloud computing isn't a panacea. I agree with those who think that major enterprises won't be headed there anytime soon (at least on the scale Microsoft or Amazon may be hoping for). Small startups like SmugMug have nothing to lose by going with S3. On the other end of the spectrum, it takes a long time to steer the
After yesterday's EC2 experience, I thought I'd find out how easy it was to replicate it on my company's private "cloud" -- i.e. VMWare and several racks. Could I make an image, store it somewhere, then fire it up anytime I needed to try something out? It turned out that it wasn't that simple. The virtual host allocation process was one of regiment. I think I'll be able to get it set up, but it might be a few days because other projects might have those processors reserved, etc.
I think my company will come around to this, but IT processes in large companies are designed to be careful and slow so as not to break things. The risk enterprises run here is not being able to cater to those who need to produce on a dime. Developers -- in my case, hackers -- are the guys who realize at the drop of a hat that they need to see what Idea X can do for them. Do they scour around to try to install everything on their machine (and risk downtime on one's desktop), try to convince their IT department to let them have a virtual machine, search around for a host to take over, or do they just fire up EC2?
This is where companies like Amazon, Google, Microsoft, whomever, get their start in remote cloud computing for the enterprise.... individuals. I find it highly suspect when people start talking about large companies' IT departments proposing a move like this.