Today Amazon announced their newest web service, SimpleDB. SimpleDB aims to bring web scale computing to the most contentious aspect of any tiered internet application, the database. Built on Erlang, SimpleDB "is a web service for running queries on structured data in real time". As a long time EC2/S3 user, I can't wait for my invite to SimpleDB (which, much like EC2 was, is currently an invite only beta).
Recently Om Malik of GigaOM wrote an article, Google's Infrastructure is its Strategic Advantage. Whie I largely agree with the premise of the article, there are two components of the infrastructure that aren't exactly touched on: people and software. Google has great people, and Google has some great software, including Bigtable, "a distributed storage system for managing structured data". Google uses Bigtable for storing most of the data for its applications (maps, pics, video, etc).
My impression with SimpleDB is that what Amazon seems to be doing is letting your proverbial David (any startup) take on Goliath (Google). Amazon isn't out to compete with Google, but frankly Google makes lots of companies nervous. By lowering the barrier to infrastructure (EC2 for compute, S3 for storage, and now SimpleDB for persistence), Amazon is allowing startups to compete with the big guys in ways that never would have been possible a few years ago.
Since SimpleDB doesn't currently support FULLTEXT search, it will be interested to see the first few libraries which enable that feature. My theory is that the first good one will look a lot like some of the techniques described in the Bigtable paper. SimpleDB also has a 1k limit for values, which I hope Amazon increases before taking the product out of Beta.
I can't wait to see what AWS releases next.
