The keynote speeches kicked off at the Expo this afternoon. Tim O’Reilly
started out by describing the latest thinking about Web 2.0. Dan Farber nicely
summarized it in this post,
quoting O’Reilly as saying: “It’s about building the global computing
network and harnessing all the collective intelligence of all the people who are
connectedĶ.We are talking about persistent computing in which we are becoming
part of a great machine”. O’Reilly’s other theme was that we ain’t seen
nothing yet on the Web – saying “we are at the visicalc stage”, in a
comparison with the PC era.
This theme of a global computing network continued when Amazon.com founder
and CEO Jeff Bezos came on stage to talk about Amazon’s latest business venture
– a giant web services platform. Bezos said that S3 now has 5 billion objects
stored in it. What’s more, it had 920 million S3 requests on its peak day, and
16,607 requests in its peak second. Obviously we’re talking about a very
well scaled system here. The term Bezos used to describe it is “web-scale
computing”.
Photo: Scott Beale / Laughing Squid
O’Reilly and Bezos then had their usual conference chat, which is becoming a
tradition at the Web 2.0 conferences. Bezos is always great value and as usual
he didn’t disappoint. When Tim asked Bezos why Amazon.com is doing platform
services now, Bezos replied at first that they are just packaging something
they’re good at and selling it. When pressed, Bezos said that for 12 years
they’ve been building a web-scale operation called Amazon.com. So what they are
doing now is building a foundation – “foundational services” he called
it – which are not making money now, but “we intend to make money”.
Nothing new in any of these answers, but the gist was obviously that long term
Amazon sees the web services platform as being a very profitable business in the
future.
Bezos also talked a little about EC2, which is their newest service released
in 2006 (Elastic Computing Cloud). Right now EC2 is invite-only, because it is
“capacity-constrained”. However Bezos said that they hope it
eventually becomes “demand-constrained” so they can roll it out to
everyone. There is no such problem scaling S3, which is already open and
available to anyone to use.
