Thursday, July 9, 2009

The cloud for dummies (or, why Google isn't taking down Microsoft)

I read about Google's announcement of their Chrome OS on the tech news sites a couple of days ago, chuckled to myself about how they seem to have lost the plot a little bit, and then went about my day. Yesterday, however, the mainstream media got wind of this. Stop the presses! Google took a shot across Microsoft's bow! This morning it was on the cover of the frickin' Financial Times for gods' sake; above the fold, no less.

Let's get this straight. Google Chrome OS (terrible name) isn't going to compete with Microsoft Windows any time soon, or ever. I won't go too deep into the details of why, though if you're interested I think Dan Lyons a.k.a. Fake Steve Jobs at The Secret Diary of Steve Jobs and Andrew Orlowski at The Register have written the most entertaining and succinct explanations of why it isn't going to make even the slightest dent in Microsoft's bottom line.

I'd like to take a moment to muse about the contrast between IT pundits and weather forecasters. Predicting the weather for the next 24 hours is a pretty exact process. Going out to five days, it gets slightly vague. Anything over a week and even the most capable forecaster will put his hands up and surrender; anything goes. With IT predictions, it's the other way around. Any techie worth his salt can describe the world we'll be living in 15-20 years from now pretty accurately. If you don't believe me, go back to the early 90s and read some stuff about how e-mail will be the primary means of communication in business and how everyone will have broadband at home - they didn't call it that yet, of course - and do their shopping online. People laughed at that, yet here we are. Go back to the same time and read predictions about what the world would be like in the late 90s, and it's all completely off.

So let me peer into my crystal ball - it's not that special, Google has one too - and tell you that 20 years from now nobody will be using computers anymore. We'll be surrounded by different size screens, cameras, touchscreens, keyboards, mice, trackballs, buttons, dials, speakers, microphones - the same input and output devices we use today, and perhaps a couple we haven't thought of yet - but instead of them being tied to a computer or a phone or a set-top box as we understand it today, they'll all be connected using insanely high-speed networking. No matter which screen you use - the huge one on your living room wall, the medium-sized one on the back of the airline seat in front of you, the tiny one in your pocket - what you'll see on it will be the same. Your e-mail, your documents, your photos, your music, your games. Your data will follow you around, accessible from any device. The only thing that will change is the size of the screen, the convenience of the full-size keyboard, the volume of the sound, etc.

You're used to thinking of a screen as an accessory to your computer, a peripheral. Your stuff - the information you create, the information you get from others and the software to process it all - is in the computer, and the screen is a portal into it. If your hard drive malfunctions or the computer goes up in smoke, it's all gone. If you want to access it somewhere else, you have to e-mail it, or upload it, or put it on a memory stick and take it with you, or burn it onto an optical disc and stick it in the mail.

Your kids, however, will think of the computer as the accessory to the screen. Their data and their software will be somewhere else. Sure, the computer will still be there, and their data will get copied and stored inside it while they're using it, but they'll be no more aware of the details of how this works than you are of which cell tower your mobile phone connects to when you make a call. Your data and your software (which, as any computer scientist will tell you, is just more data) is stored somewhere else, on the network, with backup copies all over the place so you never have to worry about another hard drive crash or leaving that important document on your home computer, and it will magically appear on whatever output device you come across, and be manipulated by any input device you happen to have handy.

People have taken to calling this amorphous, invisible computing service the cloud. It's all the rage right now, and Google is buying into it wholesale. You can see the first glimpses of the cloud already, and Google's services are prime examples. If you use GMail, your e-mail isn't on your computer; it's in the cloud, accessible from anywhere. With Google Calendar, your schedule is in the cloud. With Google Docs, your documents and spreadsheets are in the cloud.

In this world, the operating system will be very similar to Google's description of Chrome OS. All it has to do is start up and hook you up to the cloud. There is no future for Microsoft Windows as we know it in the cloud, and everybody knows it - Microsoft included.

The thing is, an this is the important bit, we're not there yet. We're not even close. Networks are too slow. Wireless coverage is pitiful. Most of our input and output devices are pretty dumb. Most of the software we use is nowhere near being ready to be used in the cloud, and in many cases the computers just aren't fast enough - sure, you can make an e-mail reader in the cloud, you can even create a very basic word processor or spreadsheet application, but if you think Adobe Photoshop or World of Warcraft, to name but a couple of examples, can be used in this model with today's technology you're just deluded. One day they will, there's no doubt about it, but not today. Not next year either, in fact probably not in the next ten years.

This isn't a new idea, and it isn't the first time people get all hyped up about the cloud - though they called it various other names back then. Back in the 90s when the Internet first came out people were going on about the exact same stuff - they called it the "thin client model" back then. Sun Microsystems even made it their corporate slogan: "The network is the computer." Looking back on this today, trying to create a cloud-based world using huge, slow desktop computers and wired dial-up connections seems laughable, and anyone who bet the farm on it back then had to pack up his stuff and move out of the farmhouse when the bubble burst.

The second bout of cloud hype that we're seeing today has come about because of the wide deployment of wireless networks and smartphones and netbooks and set-top boxes that look remarkably like fully-fledged computers that we're seeing today - people now have multiple devices and they want to access their stuff on all of them, so unlike in the '90s, the demand is actually there. But like most failed IT initiatives, Google Chrome OS and the cloud-based computing future it espouses will fail because it's long before its time. The technology and the infrastructure isn't there, and it's not going to appear overnight.

People aren't going to buy into a cloud-based computing model until it lets them do everything the desktop model does today, and that's not just e-mail and Facebook - it's video editing and 3D gaming and a bunch of other stuff too. The technology of the web is laughably unsuited for these kinds of things, and building an operating system that assumes that the web is all you need is just wishful thinking. The software architecture that this cloud-based future will be based on will be one of virtual machines, remote execution, distributed storage, massively parallel processors and numerous tiny thread-like objects being dispatched all over the place; it won't be HTTP, XML and JavaScript, for Pete's sake, and watching Google try to do it like this will one day seem as laughable as Sun's efforts in the 90s to get us to give up our desktop computers for a CRT screen and a modem seem today.

So, Microsoft isn't, and shouldn't be, worried. Windows and Office will go the way of the dodo in due course, but they know it's not time yet, and you can bet they have a strategy in place to dominate whatever replaces them. Whatever it is, you can bet the farm it won't be Chrome OS.

No comments:

Post a Comment