There were a couple of posts I came across today that got me thinking about Virtualization.
When I was putting together the Virtualization book I wrote for Intel one of the things that stood out to me is the history of virtualization technologies didn't start with VMWare. :)
It goes back quite a ways back to the IBM heydays of the 60's and 70's. Pretty much the primordial soup out of which everything else seems to have arisen in the computing landscape at least in terms of first starts.
And then the second post really hit me in terms its premise that Virtualization as a term might eventually fade into the background in the same way that few folks really understand how packets make their way across networks any more. They just jack in and off they go whether it be an EVDO, Wi-Fi, or CAT5E connection. Network packets in isolation have become as interesting as asking where the concrete comes from that makes up the interstate. Most people don't care "how" the road is made, where the white and yellow paint comes from, etc. They simply care where the highway takes them and how fast they can get to where they are going safely.
I think three years is a bit aggressive for the term to fade as I don't think we've seen all of the ways to mashup virtualization technology in interesting ways quite yet. Parallels current attempts to blur the distinction of which OS you're really running as primary with Coherence for example and other tricks like smart select, etc. are on example on the Mac OS. But I'm sure plenty of others are forthcoming for server virtualization as well.
Plus the lines between what is presentation virtualization, application virtualization, and server virtualization are just being drawn. And provisioning scenarios have yet to settle out. Plus virtualizing other pieces of the hardware stack fully such as graphics cards will create another revolution of ideas.
it might happen in 5 to 10 years but 3 years is way too aggressive. What do you think?