What’s Apple So Afraid Of?

What a bunch of “Scaredy Cats!”

That’s what went through my mind as I read today’s headline about Apple and Mac OS X virtualization. As David Marshall noted in his InfoWorld Virtualization Report blog, It seems the company is finally “loosening-up” and letting customers run OS X inside of virtual machines (previously, this was a impossible due to various clauses in the OS X EULA). However, the new capability comes with some rather draconian restrictions attached:

1. You can only virtualize Mac OS X Server. Virtualizing Mac OS X clients is still a no-no.

2. You have to run everything on Apple logoed hardware – no leveraging that Dell or HP box.

3. And, of course, you need to buy a license for each virtualized instance of OS X Server (duh!).

Restriction #3 makes some sense – if you use it, pay for it. And even #2 could be justified, if for no other reason than they’re greedy (it’s still a hardware company, after all). However, restriction #1 left me scratching my head. Why is Apple so hell-bent on keeping customers from virtualizing their Mac clients?

It certainly can’t be a profit issue. Margins for OS X sales have been pegged at 85% or more by most analysts. That’s higher than any other Apple product, including the various “i-thingies” (40-55%). With those kind of numbers, you’d think they’d want to sell as many copies as possible.

No, the only reasonable explanation for why Apple is so reluctant to let its customers virtualize OS X clients is that they’re afraid.

Apple is afraid that, if they allow virtualization, they will be tacitly endorsing the idea that the Mac OS and Mac hardware are separate entities. And once they do that – once they allow customers to start thinking about OS X running in non-Mac hardware environments (event virtual ones) – they open the floodgates to all sorts of insane requests.

Why, customers might even start demanding that they be allowed to run OS X directly on non-Apple client hardware! And if that were to happen, Apple would have to vastly expand its QA and testing programs.

It’s easy to build a fairly stable OS when your hardware deployment target is limited to a couple of dozen possible configurations (and even then, they still don’t get it right all of the time). However, when you expand that target to include the myriad permutations that dot the Intel-based PC landscape, the task of making a stable, reliable OS that performs well and meets the majority of customer requirements becomes a bit of a challenge.

Just ask Microsoft.