Ruminations on a Google Operating System

Editor's Notes: This was originally published on 11 October 2009.

The announcement came in the middle of the night, Pacific time (and hence in both my and Google’s time zone), with almost no fanfare. Google quietly declared its intentions to create a new operating system, and just like that, the dynamics of the computing world changed.

The idea is simple enough, and indeed, one that I and others have kicked around from time to time. A browser is an interesting program. In order to work cross platform, it needs to have its own graphics renderer, its own sound manager, its own file system abstraction. It needs to be able to handle events from the operating system and create its own event handler architecture to overlay that. It has to manage multiple browser tabs, a process that recently has moved into managing its own threads. When you get right down to it, a browser is a virtual operating system.

The folks at Google are no dummies. They had a browser, Chrome, one designed to do all of these things. If that browser could establish a services bridge (cf. WebMin or similar types of applications) it could readily handle all of the things that contemporary operating systems can do. What’s more, as of July 8, they also took off the little beta tags on all of their office applications - Gmail, becoming one of the most ubiquitous of mail programs, is now "official", as are the various Google Office applications. In effect, and perhaps really without fully meaning to, they created all of the core ingredients necessary for an operating system.

Now, there are a lot of things that do take place outside of those services that applications like Chrome provide. The Google announcement in fact contained another quietly understated nugget that, if its implications are fully teased out, can make your jaw drop. The Google Chrome Operating System (GCOS, until some other sobriquet is officially applied) is in fact running on top of a Linux kernel. Put another way, GCOS is in fact closer to being a Windows Manager, in effect doing the same thing as KDE or Gnome - it handles the user interface, with Linux in the back end actually managing the services. This particular design philosophy is, by now, well tested, and it makes feasible Google’s intention of having GCOS officially see the light of day by mid-2010.

Google’s intent is to target netbooks and subnotebooks first, one of the few fast growing segments in an otherwise moribund PC market, and is targeting the Intel, AMD x86 and the ARM device markets. It’s audience is those people who are most likely to use these devices - the casual web user, the student, the business traveller. Steve Ballmer must have woken up from his sleep screaming, drenched in a cold sweat.

This is not a warning shot to Microsoft across the bow. No, this particular cannon shot was aimed directly at the waterline, intending not to annoy the folks in Redmond, but to sink them. Windows 7 officially goes on sale in mid 2010, in one of the worst recessionary environments in living memory. GCOS will likely come out at worst only a few months later. Windows 7 would have likely cost around $200 for a home edition, perhaps as much as $500 for the high end enterprise edition. GCOS cost? $0 (perhaps $100 for a set of convenience DVDs and a customer support contract - the money is in the services, not the OS after all).

Getting Linux onto the average user desktop has been challenging, though it has been happening more and more. Yet this will get Linux on the desktop in a big way. Google doesn’t have OEM contracts with Dell or HP or Lenovo. It will get them. Google is all that is holding the advertising industry from the brink of oblivion - should Dell, for instance, take a pass on GCOS, Google may very well open source the OEM channels as well, which would be disastrous for the established players, and there are no shortages of vendors that would be willing to be freed of Microsoft’s fairly draconian OEM licensing contracts if there was a true alternative.

It doesn’t necessarily even put that big of a crimp on the Linux side. Google indicated that they were working with a Linux partner, which I suspect (though don’t have confirmed) is probably Ubuntu. A Gubuntu distro may very well be in the works, yet its not hard to conceive Google creating GCOS window managers for RedHat or SUSE - windows managers are useful that way.

A product announcement is, of course, not a product, Google is becoming well known for jujitsu type plays in the industry that don’t always pan out, though this one feels like it might. The public’s mood towards Microsoft is not terribly forgiving after the debacle of Vista, and while Windows 7 certainly seems to be a step in the right direction, the company has set itself up for a mass migration if the alternatives out there are sufficiently superior. Google’s media stock is higher now, by far, and just as a Chrome is continuing to take market share from both Internet Explorer and Firefox even in a fairly saturated browser market, GCOS has the potential to seriously hurt Microsoft.

This should be fun to watch.

Update