Here’s the paper I wrote for last week’s FLOSS UK Spring 2012 conference. I don’t really do slides, just scrolling through text and pointing at diagrams. But hopefully this is more useful afterwards – the paper spills all the details on how we’ve been implementing BigV, our excellent new hosting system. Have a read, and I’d be happy to answer any questions in the comments.
I’ve also dug out the paper I did for the UKUUG 2004 Bournemouth conference on what we were doing with User-Mode Linux. There’s actually plenty there that’s still relevant to planning a virtualisation system – everything from TAP interfaces to economies of scale. Both papers are a combination of half-imagined technology & business, but we did OK the first time round, so I’m thinking we have a shot with BigV!
Also at the UKUUG conference, I was fascinated to learn of Iustin Pop’s work on Ganeti, a VM supervision system developed at Google. If you’re looking to deploy a cluster of VMs on your own hardware without the bureaucracy of Openstack or Eucalyptus, I’d suggest you look at that; I certainly will be. If you do just want hugely flexible VMs with up to 8 discs and 120GiB RAM on our well-managed infrastructure, BigV is clearly where it’s at 🙂
Here are the two papers: