Serverless Computing
A cloud execution model in which the provider runs code or managed services on demand and bills only for resources used, with no servers to manage.
Serverless lowers the cost of workloads with uneven traffic, such as form handlers, webhooks and scheduled jobs, because nothing is paid while idle and no team time goes to patching or capacity planning.
A marketing site sends form submissions through API Gateway to an AWS Lambda function that validates the data, writes it to the CRM through its API and posts to Slack, handling 50 submissions on a quiet day or 5,000 during a launch.
Cold starts add latency to rarely used functions, execution time and memory are capped, and costs can exceed fixed servers at sustained high volume. Deep use of one provider's triggers and services increases lock-in.
Serverless covers Function as a Service, where single functions run in managed containers in response to events, and Backend as a Service, such as managed databases, authentication and storage. Functions are stateless and scale out per request, including down to zero. AWS Lambda, for example, bills per request plus compute time and caps standard invocations at 15 minutes. Triggers include HTTP calls, queues, file uploads and Webhook calls, which fits an Event-Driven Architecture.
Data Engineering, Intelligent Automation
Google Cloud — What is serverless computing? — https://cloud.google.com/discover/what-is-serverless-computing; AWS Documentation — What is AWS Lambda? — https://docs.aws.amazon.com/lambda/latest/dg/welcome.html
