Virtualization: Running Any OS Within OS X
While Boot Camp may have been a step in the right direction, the true holy grail for computer users everywhere is virtualization. With one company, Parallels, releasing a public beta last week and a few others— including a rumor of Apple—in the pipeline, the game is only going to get more interesting.
What is Virtualization?
There are many different meanings of the term virtual machine, but for our purpose, it is a piece of software that allows a user to run multiple guest operating systems within the main host operating system. That is, you may run instances of Windows XP, Red Hat Linux and FreeBSD that would, to the user, appear to be individual applications inside your host Mac OS X. For a slightly more technical overview, here is a recent article on virtualization.
There have been multiple solutions for achieving this on many platforms for years now. Only recently, however, has it become more feasible, more seamless and, most importantly, gained support among operating system developers. Microsoft, many Linux distributions and, as rumored, Apple are taking notice and are rolling out more solutions to co-exist.
Why would anyone want to virtualize?
Glad you asked! Boot Camp may be great, but rebooting to switch operating systems becomes tiresome very quickly. Most users will end up sticking with one operating system, switching only when they really have to. With virtualization, it all becomes seamless. Also, many companies and people have legacy software- that is, software that works well only with older operating systems- that they just can’t let go. Now they can run an older operating system like Windows 3.1 under their Mac OS X and all their software can live happily ever after.
Virtualization is also extremely beneficial to application developers who can test their software with different versions of operating systems or applications painlessly.
And even for the rest of us, who just want to use a few applications, that annoying browser for those arrogant web sites and those glorious Windows games, virtualization is the way of the future. A future where the device and the network are the platform—not the operating system.
All of this has become especially important for Macintosh users since the move to Intel based machines. Virtualization is a lot easier when the guest and host operating system were built for the same CPU architecture. For the virtual machine developers of the world, the Mac OS X just became much more interesting.
The Players
As usual, the open source community was the first out of the gate with a working solution called Q for people looking for virtualization on the Intel Macs. With Microsoft still not ready with Virtual PC for Intel, this market is getting larger as more people look to migrate to new Macintosh computers.
Q is an open source cocoa port of QEmu which is an older x86 emulator that allows you to install any x86 based operating system on your host OS. Q allows that host OS to be Mac OS X Tiger. It can even import old Virtual PC disk images for use in Q. Best of all, it’s free- but still in a very early alpha stage.
VMWare, the big player in the virtual machine market, is still to commit to an OS X version of their popular software. A play by VMWare would be very interesting as they have the enterprise marketshare and support structure to make it feasible for corporations. VMWare has claimed, however, that they have gotten it to work in their labs.
The big rumored player, of course, is Apple itself. As I mentioned previously, the rumors in the Mac community speak of Mac OS X Leopard—the next version of OSX- with built in support for operating system virtualization. The arguments for and against such a move are the same as they are for Boot Camp, except that now there are no excuses to ever actually boot. This provides you a loop hole out of the Windows no-booting pledge here at Apple Matters. This is probably the best argument for virtualization yet!
And finally, there is the newest and most impressive kid on the block—Parallels.
Last week, Parallels released a beta version of its virtualization solution for Intel-powered Macs. This version is currently available for free, and with future full versions priced at under $50, they clearly outdo VirtualPC. Not only that, but if this software works as advertised, it may even make Boot Camp unnecessary. By supporting any x86 based operating system, from Windows to Linux to FreeBSD to Solaris, Parallels does a lot and what’s more, it already has a working solution. Also, Parallels has the advantage of being a company with a short but proven track record in providing cheap commercial virtualization solutions to every kind of customer. It is precisely the kind of company that could push mainstream adoption of virtualization.
It appears that the beta does work quite well—including loading Windows XP in under 10 seconds! The company also has screenshots of things you would never have though you would see in this lifetime such as Windows 3.1 and Solaris running simultaneously under Mac OS X.
Of course, the big fear for any of these companies is that Apple will roll this into the next version of OS X and any third-party development will become redundant unless it is a lot better. In either case, however, the landscape for the consumer can only get more interesting.
del.icio.us
Digg
Facebook
Slashdot
StumbleUpon
Comments
I think Apple was holding BC back until someone figured out how to get XP running on an Intel Mac. The timing was too perfect. I also think that MS would be working furiously on the Intel VPC update, just because of the new market possibilities for them. I also think that with the way MS has been dropping their Apple products lately, only commiting to support Office for Mac for another 5 years, plus their track record for new products, Apple would be silly to count on MS for virtualization. This is one of Apples best chances for new growth. Superior HW & OS didn’t work. The iPod did. So if Apple turns themselves into a consumer gadget company that sells computers as a sideline, they may have some success, but they will have to expand their line of products & compete directly with Sony.
And where is Apple’s compelling enough reason to legally allow that? So far the counter-arguments have been and remain more convincing to me.
I don’t have any inside info, however as a long-time IS person and Apple watcher, I think:
(1) Apple probably released Boot Camp a bit sooner than they might have because of the hype and the contest to find a way to do it. Apple realized that it benefits them to have a surer, “Apple-approved” way of doing it.
(2) Leopard will nearly certainly NOT have virtualization built in. Apple is going for a production version of Boot Camp as integral to the OS. I don’t think Apple is ready for virtualization yet. It does take a considerable amount of thought and time to develop that technology. So, if Parallel gets it to market in a stable product, they should have several years of good revenue from it before Apple attempts to repeat it. Indeed, Apple may not even want to do that at all—at least initially—and decide to leave that functionality entirely to them.
And where is Apple’s compelling enough reason to legally allow that?
The only real counter-argument is that Apple hasn’t decided to do it yet and therefore it must be right not to.
But the reasons for licensing OS X are manyfold. Unlike Boot Camp, that still requires a switcher to pay Apple’s premium prices for hardware to “try out” a Mac, buying a stand alone version of OS X for a couple hundred bucks would be much more feasible for most users looking to try something different. Look at the inroads Linux has made.
It would be much more profitable for Apple. While hardware sales would likely not be that affected, any losses in revenue would be more than offset by the almost pure profit margins of selling OS X in a box.
It would get OS X into the hands of manufacturers like Dell and Gateway, who would love the opportunity, and that kind of marketing muscle would be invaluable for the Mac. And most of the driver support would come from the hardware maker, not Apple.
The only real valid reason I’ve heard not to do it is the hardware driver issue. But that seems to me a remarkable lack of confidence in OS X’s ability to handle different kinds of hardware other than what Apple sells you in a box. I think that as well as it handles third-party add-ons like external drives and printers, that a gfx card shouldn’t be that much of an issue. And Apple could certainly recommend hardware specs.
Aw shucks, I hoped you’d come up with a reason I hadn’t heard for why it’s in Apple’s best interest to license OS X instead of repeating ones that have already been exhaustively discussed.
Some counter-arguments I generally agree with are elaborated under “Why Apple Isn’t Going to Release Mac OS X for Other PCs or License the Mac to Other PC Makers” in John Gruber’s most recent Boot Camp article. [please spare us any personal conjecture about him this time?]
The main point of differentiation for me isn’t whether Apple should/shouldn’t but rather why they would/wouldn’t. Like anyone else, I can think of “should” reasons but those remain stubbornly unconvincing in light of several “wouldn’t” rebuttals.
Obviously it could eventually happen but seriously considering it as a near-term reality seems frivolously premature. Who honestly thinks it would occur before Apple finishes transitioning the Mac product line to Intel? Speculations on the topic could become more intriguing after that.
If it’s a question of will they or won’t they, Apple fanboys (like daringfireball) are notoriously bad prognosticators. Remember when Macs would never boot Windows and all of the “compelling” reasons why? Note daringfireball’s “shock” at the announcement of Boot Camp. That’s because their judgement is based almost entirely on what Jobs has said or what they surmise as Apple’s philosophy. Which is why his reaction so quickly morphs into total acceptance of the “obviousness” of the decision - if Apple did it then it must be right.
But what Jobs says is bullsh-t most of the time, and Apple has no real philosophy beyond making money. So no one can provide any legitimate argument why they won’t license OS X because whether they will or won’t is complete speculation.
But we can, if you’re truly interested (and it doesn’t seem like you are), discuss why they should or shouldn’t. And I’ve listed some of the reasons why I think they SHOULD above.
The arguments for why they shouldn’t, at least those provided by daringfireball, aren’t any more compelling (far less, actually). And he trips over himself constantly, from point to point, trying to keep up with a dizzying array of contradictory talking points.
The funniest one is when he insists that licensing would be disasterous for Apple because even some existing Mac users wouldn’t buy Apple hardware because it’s more expensive, even though that’s impossible because it isn’t more expensive at all.
If Apple hardware is truly competitively priced, then it’s difficult to predicate your argument on the belief that people won’t buy it for a reason that you think doesn’t exist.
The rest of it follows along those lines: Even though users will pick OS X over Windows given a choice, they shouldn’t actually be given that choice in case they don’t pick it.