Amazon Puts out the Storage Infrastructure Welcome Mat
- By John K. Waters
- March 14, 2006
Web Services (AWS) has just released a simple API that can be used to store
data in, and retrieve it from, the company's own storage infrastructure.
Unveiled today at SD West, Amazon S3 is
one of those slap-your-forehead releases. I mean, who better than Amazon--an
online retailer who's very existence depends on data storage that is fast,
scaleable, and reliable-to-the-nth-degree--to offer Web-based data storage.
''Companies like Amazon can afford to put years of effort and resources
into developing management software that can be smart about running on cheap
hardware and tolerating failures and that sort of thing,'' says AWS VP of
product management and developer relations Adam Selipsky. ''And they can afford
to spend millions to buy lots of that hardware so that they get the benefits of
scale and reliability. Needless to say, most developers don’t have that kind of
time or money.''
Amazon S3 was intentionally designed with a simple feature set, Selipsky says, so
that it will work well with any type of application. Among those on the short
list of S3 beta users are the University of California at Berkeley's [email protected] team, which is using Amazon S3 to
store and deliver 60,000 images collected in its ''aerogel'' experiment to
100,000 volunteers, who scan the images looking for dust particles from Comet
Wild2; and the castingWords podcast
transcription service, which uses S3 to store and retrieve original audio files
and the transcribed texts.
S3 users are charged 15 cents per gigabyte of data stored, and 20 cents per
gigabyte of data transferred. There is no minimum fee, and users pay only for
what they consume.
''This is Amazon again extending its portfolio by adding services for selling
infrastructure capabilities,'' says ZapThink senior analyst Ron Schmelzer.
''They are basically saying, we’ve built this platform over the past 10 years to
provide highly reliable and scalable delivery of electronic assets. And, hey, we
can monetize it by providing access to developers.''
Amazon isn't the only one thinking along these lines. It has been reported
that Google is interested in developing a similar product. Schmelzer sees the S3
release as part of an overall trend among the major Web app providers--Google,
eBay, Amazon, MSN, and Yahoo--to reach beyond their original business models.
''They're all expanding their footprints,'' he says. ''MSN isn't just an ISP; Google
isn't just a search company; eBay is more than an online auction house;
and Yahoo is more than an online directory.''
S3 also represents yet another change in the way people
are building applications. The infrastructure demands of complex apps aren't
going away, but, by allowing developers to leverage powerful, existing
infrastructure capabilities, companies like Amazon are making it possible to do
more with less.
Amazon S3 is
available today at http://aws.amazon.com/s3.
John K. Waters is a freelance writer based in Silicon Valley. He can be reached
at [email protected].