Lately there'd been a bit of tech-blog speculation that Google would do this, and then it went and did it: Last night Google opened up its data centers to developers outside the company.
Google's own programs run on hordes of computers spread out in warehouses around the world. The system, strung together with home-brew code, is shrouded in secrecy, and is thus legendary -- people outside the firm aren't sure what Google's machines can really do, but you sometimes hear suspicion that Google might possess nearly supernatural supercomputing capacity.
Google had long kept all this to itself. Now, through the Google App Engine, the company will give start-up Web firms access to the beast.
If you've got a hot new idea for an online app -- say, a nifty calendar or a revolutionary e-mail system -- you no longer need, now, to rustle up a Web hosting plan, set up a database, and buy up servers as your traffic grows. Now you can turn the whole thing over to Google, which will manage your traffic and computing needs on its own machines. Think of it as Henry Ford leasing out his assembly line to aspiring manufacturers during off-hours.
But wait a second. Why would Ford do that? Google's data centers, like Ford's assembly line, are a competitive advantage -- why give that up to a potential rival?
There are two good reasons: First, because there's money to be made. Google will offer a bit of computing power for free, but companies will have to pay for more resources. Web hosting is a big business, and Google is the most powerful Web host in the world. Why not sell some of that power?
But Dave Winer hit on a more important idea: The App Engine lets Google keep its rivals close, because today's rival is tomorrow's acquisition target. By giving start-ups access to Google's servers, Google is encouraging a class of developers to create programs according to Google's specifications. If a few of these start-ups succeed, Google can snap them up and integrate them into its operations with no cost -- after all, their programs were built to run on Google.
Several years ago Amazon.com began unveiling a slate of services similar to those Google is now providing. First, the retailer offered online storage to start-ups, then computing power, a database, and even a "fulfillment system" (in which the company will process and ship physical items for other companies).
There'd been talk of the Google system being an Amazon-killer, but the two initiatives are technically quite different.
Amazon has no free option; you pay for everything. Amazon, though, sells its system in pieces -- you can buy online storage by the gigabyte, computing power by the processor-cycle, in whatever configuration you choose -- while Google will sell you only the full app system, hosting and serving the entire program by itself.
Google will also only calculate data in response to Web queries from users of your app -- its system won't, say, do an hour-long computationally intensive analysis of all the numbers in your database.
Depending on what kind of start-up you're building, you might favor one approach more than the other. Zonbu, a firm that sells people Linux computers and online storage for a monthly fee, uses Amazon's system to keep its customers' data. Because it's not a Web application and doesn't require computing power, Zonbu couldn't use Google's App Engine.
Neither could you use Google's system to build a rival search engine: Because the App Engine only processes data in response to users' queries, it wouldn't be able to crawl and index the Web, a necessary part of creating a search engine.
The vast majority of Web start-ups, though, would benefit from Google's all-in-one approach. I'll defer again to Dave Winer -- the emphasis here is his:
What you're going to see here that you've never seen before is shrinkwrap net apps that scale that can be deployed by civilians. That's a mouthful, but that's what's coming. Why? Because here is a standardized platform that can be stamped out in the billions of units. Maybe Google can't do it, but the perception is that they can. Who is willing to stand up and say Google hasn't nailed scaling? What PCs did in the 80s, Google is doing now. PCs took the black magic out of owning a computer. Now Google is taking the black magic out of operating a scalable web app.
That's the story here: What Google is building is something like the next PC. It's a foundational system, sitting beneath a cornucopia of new Web apps. This could be huge.