Silliness aside, picking the best model candidates is one of Mike's key responsibilities: evaluating existing computing paradigms and making predictions on which computing models will succeed. As you may guess, Intel has a vested interest in staying at the top of the game.
We didn't directly discuss Intel's Emerging Compute Models & Their Status in the Market study, released last month. Our extended interview today was intended to cover the current challenges of VM management in the enterprise. In case my editors or his boss are reading this -- mission accomplished. Management issues were discussed, the state of the market was reviewed, and basic strategies for success were identified.
Mike and I also went down a number of rabbit holes. Deep, dark, twisty ones looking at possible futures where every computer (server, desktop, laptop, thin client, or hand-held gizmo) is a VM platform. Where it makes sense to host single-user, single-OS boxes on top of an integrated hypervisor ... where purpose-built desktops ship with locked-down virtual security and management appliances unbound from a user's preferred OS.
I'll go so far (I won't put Mike on the spot here) that chip vendors become almost irrelevant in a world where all operating system instances are abstracted; if a VMM is managing hardware calls, IT admins or OEMs can build images for their VM host platform of choice without worry over motherboard chipsets or CPUs. The fine tradition of building an operating system image to fit a specific production run of workstations will become a quaint old-timers' story, right up there with punch cards, six-packs under the raised floor, and OS install discs.