On November 13, 2014, Amazon Web Services announced a preview of AWS Lambda. In the launch post “AWS Lambda - Run Code in the Cloud,” Jeff Barr described it as “a brand-new way to build and run applications in the cloud,” one that requires no server management or capacity planning. Lambda is a zero-administration compute platform: developers upload code, and AWS runs it on demand.
Lambda is event-driven. The launch post explained that the service would automatically run code in response to modifications to objects in Amazon S3 buckets, messages arriving in Amazon Kinesis streams, or table updates in Amazon DynamoDB. Instead of provisioning and running servers continuously, a developer registers a function and the events that should trigger it.
The billing model was a defining feature. The announcement stated that customers pay for compute time in units of 100 milliseconds and pay for each request, with a free tier of one million requests per month plus compute time that varied with the memory allocated to each function. There was no charge for idle capacity, because there was no idle capacity to pay for.
The AWS Lambda product page continues to frame the service the same way: run code “without thinking about servers or clusters,” with AWS handling all infrastructure management, automatic scaling up and down based on real-time demand, and “pay-per-use billing by the millisecond.” Lambda’s success made the function the unit of deployment and helped popularize the broader serverless model.