The big News seems to be that RedHat and AMD managed to do Live migration of a virtual machine between 2 different CPU Vendors.
Now given my age and my starting alzheimer, LinuxKongress 2005 in Hamburg seems like ages ago,
As a speaker giving my Automating Xen Deployments talk early on the first conference day , I flew in the day before the conference and I needed some network connectivity so I sneeked into Ian Pratt’s tutorial about Xen. I remember sitting between Heinz Mauelshagen and my colegue Peter Leemans, Heinz was playing with SMP guests on a non SMP host and tried to figure out the limits.
If I recall correctly at the end of his session Ian was explaining Live Migration and demoing it by migrating a small virtual machine around between laptops of people in the front rows of the audience , Laptops from different architectures and from there started a discussion about CPU feature checking.
I asked around and some people actually remember parts of the discussion and the demo.
So for me , when RedHat and AMD claim today they have achieved in something new that was Never going to be implement they are either focusing on nitty details, like migrating from 2 brand new CPU’s to each other and forgetting the fundaments already existed for ages or just trying to get some positive news out of the door. In which they succeeded.
But I`m looking out for the next step in virtualization , not just mashups of things we’ve all been doing before, or things that are really similar to existing things.