The Death of the Desktop?

Posted November 20, 2009

Google today revealed details of its operating system: Chrome OS. There is a reason that they named their operating system after their browser: their operating system IS a browser. Your computer boots in seconds and up comes a browser.

Take more than a cursory look at Chrome OS and you realize that Google is offering a vision for the next era of computing. They answer a lot of questions but raise even more.

Dude, where is my hard disk?

Gone. Chrome OS only works on devices with flash memory but no hard disk. You can have local files and the OS automatically encrypts them. Otherwise your C: drive is now the Cloud.

Without applications or data, your computer is a lot less important. A Chrome laptop with a full screen and keyboard would cost $3-400 and if you lost it, you wouldn't lose any information. Replace it and your apps and data will be waiting for you. A thief would get no information of yours.

Um, what happens to my local applications?

They went to heaven. Chrome runs all apps from the cloud. The first tab of the browser contains a list of apps: mail, documents, social networks, shopping sites, games, etc.– all online. Any local files are instantly backed up on to the big hard disk in the sky — the one that is configured so that even when one of its zillions of hard drives fail, your data is not affected.

What about viruses?

Chrome OS treats all applications as hostile, so if an online application gets infected or misbehaves, it cannot affect or infect anything on the local drive. And it is much harder to infect a well-configured online application.

When infections occur (and they will), they are managed centrally. You do nothing — just like an online application like Google Docs is today. You never download annoying updates — but the apps may get ten updates a day. As a user, all you know is that it keeps getting new features and working better.

What about device drivers for printers and cameras, etc.?

Chrome OS will need drivers and a developer community to fill in the gaps. This is not a trivial matter. At present, browsers can barely talk to devices. Try printing a Google document — it's pathetic.

What if I am not connected?

This OS wants you to be connected. We can assume that Chrome OS will feature Google Gears, the current solution for running Google docs offline. At its best, Gears makes it seem like you are connected (at worst, Gears simply doesn't work). With drivers, connectivity is the soft spot. Google is assuming a 24/7 wifi connection, which many people have. Those without it however, cannot move to Chrome OS without a lot of pain.

Where does this leave Microsoft?

MS has an operating system business that will be around for a long time. I would not be surprised if millions of people are running Windows 7 in ten years (it's a strong OS). It will be a legacy, not a growth business however, because fundamentally, nobody wants an operating system. We want our apps and our data — the rest is overhead we would rather avoid. Who wants to pay $200 for something you don't even want? Seems likely that Microsoft will see it's Windows desktop franchise go away about as fast as it's enterprise server operating system business did when Linux came out.

Microsoft also makes applications — notably Office. They have moved it online, where early indications are that it is a very strong web app. The problem for Microsoft is how to charge money for web apps. They will have to offer a free, ad-supported version. Companies and consumers will pay something to have the ads go away — but probably not hundreds of dollars as we pay now.

Why didn't Google just extend Android? Or did they?

They did not. Chrome OS is it's own beast and is being released to developers as an open source OS like Linux. How or if it fits with Android, their Mobile OS, is unclear to me.

Google looks at the world of mobile apps and desktop apps as fundamentally and perpetually different. The case for this is that Android needs to run on hundreds of devices with different screen configurations and needs to comply with myriad carrier requirements. To the extent this becomes less constraining, a mobile version of Chrome OS would replace Android at some point.

Apple, of course, did this. They built the OS for the iPhone on a stripped
down version of OS X, their version of Linux for the desktop. This has big stability advantages if you build desktop operating systems. Google isn't building a desktop OS — it wants to transcend them. For this reason, Apple isn't much happier
than Microsoft about Google's effort to push apps to the cloud — the latest squeeze on Apple from Google.

Will Google make notebooks? Or phones?

The notion that Google will release a branded phone, supported by data only subscriptions from ATT and VOIP integrated with Google Talk is a hot rumor at the moment.

But hardware is WAY outside Google's area of competence. They may want to demo the future as a way to catalyze handset makers and carriers to try something new. The rumors seem odd to me, although they are getting a lot of verification by tech journalists. One possibility is that the Google phone is not really a phone at all — although you can make phone calls on it. It would be a cheap handset that runs Android built around Google Talk. The devices would require WiFi.

If Google made them cheap enough, G-phones would increase demand for public WiFi and would promote Android. Since ATT and Verizon are ramping 3G capacity at an astonishing rate and now require that handsets be WiFi-enabled to help relieve network load, this would not be competitive with carriers or even with handset makers.

Would anybody use a G-phone? Suppose it cost $75, requires no contract and carries a SIM-chip for pre-paid phone cards? Prepaid phones are huge everywhere in the world except the United States. You could point Google Talk to the phone which, with a full suite of Android apps, would be a real Iphone competitor and would create a huge new market.

So does cloud-computing spell the end of the PC?

Yeah — it does. Google is betting that the grand computing architecture will continue to evolve, driven as usual by the cost of processing, storage, and bandwidth. When all were expensive, we had mainframes. As processing and storage got cheaper, we got minicomputers and desktops. As bandwidth got cheaper, we networked everything and the internet took off. (This was not obvious in the late eighties. I recall a room full of McKinsey partners nodding their heads soberly at the proposition that there was "no business case for networking desktop computers").

Once the marginal cost of storage, bandwidth, and processing drop to near zero, devices like netbooks become cheap and ubiquitous and Sun's memorable maxim that "the network IS the computer" starts to make sense because hosting apps and files in the cloud is feasible and has real advantages.

Desktops as we know them won't all die, but they will become a lot less important. Google is betting that we will all use netbooks with full screens, flash drives, and keyboards or tablets. Sure, a few senescent boomers will declare this a return to early 80's, but it's not. Hooking a dumb terminal to a corporate mainframe and hooking Chrome OS to the internet are not remotely the same thing.

Why is it called "Chrome"?

Chrome is the frame of a browser — the trim along the edge. The term is used by graphic user interface designers. It is also element 24 in the periodic table, known for its corrosion resistance and toughness. Not to be confused with Silverlight, Microsoft's product that may compete with Chrome. It is named after element number 47 and known mainly for its ability to conduct heat and bedazzle explorers.

Who wins and who loses?

If Google's vision prevails over a ten year period, everybody wins except for Microsoft, utility makers, and phone companies.The unknowns are Apple and Intel.

Flash memory makers win, but disk drive makers also win ("the cloud" is actually buildings full of commodity hard drives). Applications win, utilities lose (who really wants to defrag their hard drive or update antivirus software?). Apple, which already hosts more than 100,000 web apps for the iPhone, probably adapts fine and might turn out to be better at implementing this vision than Google is.

Intel may lose if chips optimized for browsers become commodities and the "Wintel" monopoly of chips built to run vital desktop operating systems disappears. Wifi wins and public wifi really wins (and this is a market Google may enter in a serious way, since they own a chunk of spectrum. It's a huge market that Intel, among others, can benefit from). Traditional telephony loses (again).

The biggest winner by far: consumers.