Serverless computing
Serverless computing
Serverless computing is a cloud computing execution model where the cloud provider dynamically manages the allocation and provisioning of servers. The term serverless does not mean that no servers are involved; rather, it allows developers to build and run applications without having to manage the underlying infrastructure. Code is executed in response to events (such as HTTP requests or database changes) and is typically charged based on the exact amount of resources consumed, rather than on pre‑purchased capacity.
The model is most often associated with 0 (FaaS) platforms, but also includes 1 (BaaS) offerings that provide managed storage, authentication, and other services. Serverless architectures are valued for their operational simplicity, automatic scaling, and reduced cost for intermittent or unpredictable workloads.
History
The concept of serverless computing emerged in the early 2010s as cloud providers sought to abstract infrastructure further. In 2014 Amazon Web Services (AWS) launched AWS Lambda, the first widely adopted FaaS platform, enabling developers to run code without provisioning or managing servers. Google followed with Google Cloud Functions in 2017, and Microsoft released Azure Functions in 2016. The model gained popularity alongside event‑driven architecture and microservices.
Earlier precursors included services like Google App Engine (2008) and Heroku (2007), which abstracted operations but still required configuring runtime environments. True serverless removes all server management from the developer's view.
Features
- No server management: Developers upload code and define events; the provider handles scaling, patching, and availability.
- Automatic scaling: Instances are created or terminated in response to demand, often from zero to thousands of concurrent executions.
- Pay‑per‑use pricing: Charges are based on execution time and memory consumed, with no cost for idle resources.
- Event‑driven execution: Functions are triggered by events such as HTTP requests, file uploads, database updates, or message queue messages.
- Stateless functions: By default, functions run in a stateless container; persistent state must be stored externally (e.g., in Amazon S3 or Amazon DynamoDB).
Limitations
- Cold starts: When a function is invoked after a period of inactivity, the provider must allocate a new container, causing latency.
- Vendor lock‑in: Serverless platforms often rely on proprietary APIs and services, making migration difficult.
- Limited execution duration: Most providers impose a maximum timeout (e.g., 15 minutes on AWS Lambda), unsuitable for long‑running tasks.
- Debugging complexity: Traditional logging and monitoring tools may be insufficient; distributed tracing is often required.
Use cases
Serverless computing is commonly used for web application backends, API endpoints, real‑time file processing, data transformation pipelines, and IoT event handling. It also suits scheduled tasks, chatbots, and stream processing.
Providers
Major providers include AWS Lambda, Google Cloud Functions, Azure Functions, Cloudflare Workers, and IBM Cloud Functions. Open‑source alternatives such as Apache OpenWhisk and Knative enable self‑hosted serverless platforms.