Amazon Simple Storage Service (S3) was announced on the AWS blog on March 14, 2006, by Jeff Barr. The launch post described it as “reliable, highly scalable, low-latency data storage” accessible over the web through both SOAP and REST interfaces, marking one of the first widely available pay-per-use cloud services.
S3 organizes data as objects identified by user-defined keys, with associated metadata and Access Control Lists that govern privacy and sharing. At launch, individual data blocks could be up to 5 GB, and the service stored copies across multiple locations to provide high data availability. Its pricing was deliberately simple: by the launch post, storing 1 GB of data for 1 month cost 15 cents, with data transfer priced per gigabyte.
The importance of S3 is that it turned storage into a metered utility. Developers no longer had to buy and operate disk arrays to hold growing volumes of data; they could store and retrieve objects on demand and pay only for what they used. This made S3 foundational cloud plumbing, used to hold everything from website assets and backups to the machine images that boot Amazon’s own compute service.
S3 launched alongside the broader Amazon Web Services effort and, together with EC2 later the same year, defined the practical beginning of the public cloud.