The year 2006 marks the practical beginning of the public-cloud era, when Amazon Web Services launched two services that together defined the category. The first was Amazon Simple Storage Service (S3), announced on the AWS blog on March 14, 2006, as “reliable, highly scalable, low-latency data storage” priced at 15 cents per GB-month.
The second was Amazon Elastic Compute Cloud (EC2), announced on the AWS blog on August 25, 2006, as a limited beta. EC2 offered rentable virtual servers built on the Xen hypervisor and billed at 10 cents per hour, booting from Amazon Machine Images stored in S3.
What made 2006 a turning point was the pairing. S3 supplied durable, pay-per-use storage and EC2 supplied on-demand, hourly-billed compute, both accessible over the web through published interfaces. For the first time, a developer could assemble a complete production infrastructure without buying a single piece of hardware, paying only for the storage and server-hours actually consumed.
This shift from owned capital assets to metered, rentable infrastructure is what people came to mean by “the cloud.” The 2006 launches set the template that competitors and the wider industry would follow for the next two decades.