Making the Web Safe for Computers
The strategy behind Amazon Web Services is to give programmers virtually unlimited access to the very foundation of Amazon’s business–its product database–whether they are inside or outside the company’s walls. Developers can grab product data, reformat it, add related services, and use it to attract eyeballs to their own sites. If they feel like it, they’re even free, like Taylor, to create parallel-universe Amazons that have the added features they crave. Amazon demands only one thing in return: that visitors to these satellite sites complete any purchases through Amazon.com itself. The site owners, meanwhile, earn a decent commission on each sale.
Exposing the world’s largest product database–along with the editorial content and personalization functions that make Amazon.com so uncannily useful–is such a counterintuitive business strategy* that analogies are hard to come by. In a way, the risk the company is taking is like the one Apple Computer has always avoided by refusing to license its Macintosh operating system to other manufacturers. Some observers still think Apple missed its golden opportunity to mount a comeback against Microsoft, while others maintain that the Mac OS is Apple, and that putting it on other machines would have diluted the brand. In Amazon’s case, outside programmers could find cleverer ways of using Amazon’s data than Amazon itself and end up sucking away so much traffic that Amazon’s own site cedes e-retailing’s center stage.
And yet Amazon’s move also reflects the spirit of the age–or at least, the spirit among the Web’s technical avant-garde. As Amazon CEO Jeff Bezos put it in a speech at October’s Web 2.0 conference in San Francisco, “Web 1.0 was making the Internet for people; Web 2.0 is making the Internet better for computers.”
Web services, whose name became a marketing meme years before the technology to make them work had actually matured, boil down to the simple but powerful idea that the Web should be more than just a way for human users to call up preformatted documents. It can also be a medium for software programs to communicate and share data with one another. Through a nonproprietary formatting scheme called the Extensible Markup Language, or XML, a string of data can be labeled according to type–as, say, a phone number, a price, or a book title. Web software can then harvest the data from a remote site and re-present it to the end user in any way a programmer wishes. That means a company with a trove of data it wants to share need only put it on the public Web and give programmers a few simple tools for accessing it.
Though these tools have a geeky and daunting name–“application programming interfaces,” or APIs–their function is easy to understand: they are short hooks of computer code that allow smaller programs to communicate with large software systems, such as Microsoft Windows. In the case of Web services, an API amounts to a set of commands programmers can use to interact with large, database-driven Web sites.
For a company that makes its APIs public, the real trick is figuring out how it will be rewarded. And in this respect, too, Amazon has done exceedingly well, signing up thousands of programmers who are sending it millions of hits a day, though the company won’t disclose how much it’s earning through their satellite sites.
There isn’t one tutelary genius responsible for Web services’ ascent at Amazon. Even Bezos was reportedly awakened to their possibilities only after a bit of prodding from friendly outsiders.
Rather, this is a tale of clever Amazon software engineers and marketing professionals who were attuned to emerging technologies like XML, the Simple Object Access Protocol (SOAP), and the Semantic Web, and who saw how their efforts in the 1990s to increase Amazon.com’s accessibility led inexorably to the Web services model. Increasingly, it’s also the tale of tens of thousands of outside Web developers with a fixation on new technologies and an urge–with Amazon’s blessing–to see just how far they can stretch that model.