Amazon in the hosted storage business–kind of
I just learned about Amazon’s new Simple Storage System, or S3. Amazon are clearly getting serious about hosting key tech plumbing, having already offered its simple queuing service and Mechanical Turk, which I expected to be much more popular than it is.
Afaik, the approach Amazon are taking with S3 is somewhat unique. They are providing a simple REST and SOAP API, built around a lightweight object idiom, where an object is identified by a developer-defined key, and consists of one byte to five gigabytes of data. Access is via HTTP/1.1 and Bittorrent. Object access control is user-based, or can be made public.
Pricing is interesting too; $0.15/GB/month, plus $0.20/GB transferred. No setup fees, minimum purchases, contracts, or bullshit.
It’ll be interesting to see who does what w/ this service. Amazon are supposedly using it internally for all of their own data storage needs, but apart from that I can imagine some compelling third-party applications built atop the Amazon storage infrastructure. In effect, a small team of resource-constrained developers could leverage effectively limitless storage and bandwidth resources with very little upfront cost, which is a substantial qualitative shift in tech infrastructure economics. Then again, I was similarly agog over Mechanical Turk, and that’s gone nowhere…
Tags: amazon, Migrated from Drupal, s3, tech diary