Virtualization - A cold reality for the Standard OS
Lots of people and companies are excited about Intel's Vanderpool and AMD's Pacifica virtualization technologies. You hear constant clanging about "higher utilization" in the datacenter and the "cool" factor associated with being able to run Windows on a Mac (which is kind of cool, because you get OSX plus all the apps monopoly power can buy on a beautiful machine). But has anyone stopped to consider what widespread virtualization might do to break the innovation stranglehold created by the "standard OS disease" (SOD)? Most companies and consumers are not even aware they have SOD because they have learned to live with the technical equivalent of chronic fatigue disorder.
SOD is the condition of settling for slow innovation and poor performance while shelling out extra money to make things work together because all new innovations must be filtered through the least common denominator of a "standard OS." The current tonic for SOD is expensive and bitter. For commercial concerns, the tonic for SOD is expensive system techs that constantly fight through "patch hell" to pull together a rack of machines, all running variants of one or more "approved, standard OSes", that collectively deliver the value of some enterprise application. For consumers the SOD tonic has 2 flavors. Flavor one is a semi-annual re-install of Windows to recover from the bloat created by the helter skelter of conflicting application dlls and Windows updates. Flavor two is shelling out the dough to own a couple of machines, with the Windows machine relegated to only the chores that your Mac or Linux box cannot accomplish because some website or app you require only supports Windows (monopoly power has its rewards).
Ultimately, widespread and high performing virtualization is the best tonic for SOD because the application provider can simply deliver to the user their application PLUS whatever fractional OS is required to run the application most effectively. The base OS that accesses your hardware (domO for the techies) is NEVER (or rarely, in the case of a security patch) modified for the life of the machine, and applications run seamlessly together because they simply plug into "guest" sockets on the base OS.
Why does this new virtualization tonic break the innovation stranglehold? Because application vendors no longer need to spend 30 - 40% of their engineering time and money "porting" and supporting the application for all of the variants of "standard OSes" (versions, patch levels, etc) that might be in use by their target market. This time and money can be re-directed to features and performance - things that matter to customers. Beyond getting better applications faster, customers also jettison the bitter tonic of extra machines and extra time currently required to battle SOD.
What happens to the OS vendors? I suppose it depends on their reaction to the new tonic. If they are allergic to the notion of the application provider streamlining their "standard" to better suit the needs of the application, they are in for a rough ride. If they are "flexible" (a rarity for system software vendors), they will probably pick up market share at the expense of their rivals. It should be fun to watch . . .