Serverless Computing

Definition

Serverless computing is a cloud execution model in which the cloud provider dynamically allocates and manages the underlying servers, automatically scaling resources up or down (including to zero) based on demand, so developers deploy code without provisioning or managing infrastructure directly. Billing is typically based on actual execution time and resources consumed rather than pre-allocated capacity.

Overview

Despite the name, serverless computing still runs on physical servers — the term refers to the fact that developers and operators no longer need to provision, patch, or manage those servers themselves. The cloud provider handles allocating compute capacity on demand, scaling it automatically in response to incoming traffic (including scaling down to zero when there's no traffic, meaning no cost during idle periods), and abstracting away operating system maintenance, capacity planning, and load balancing that would otherwise be the operator's responsibility in a traditional server or even a typical container-orchestration setup. The most common concrete implementation of serverless computing is Function as a Service (FaaS) — platforms like AWS Lambda, Google Cloud Functions, and Azure Functions, where a developer deploys individual functions triggered by events (an HTTP request, a file upload, a queue message) and pays only for the actual compute time consumed per invocation, typically billed in sub-second increments. But 'serverless' as a broader category also includes serverless databases (like DynamoDB, Aurora Serverless, or PlanetScale), serverless container platforms (like AWS Fargate or Google Cloud Run, which run containers without managing the underlying VM fleet), and serverless message queues and storage services that scale automatically without capacity planning. The trade-offs of serverless computing are well understood: it removes operational burden and can be extremely cost-efficient for spiky or unpredictable workloads (since idle capacity costs nothing), but introduces its own challenges — cold-start latency when a function hasn't run recently and needs to initialize, execution time limits imposed by the provider, more complex local development and debugging workflows, vendor lock-in around provider-specific APIs and event formats, and a shift in cost model that can become expensive at very high, sustained, predictable traffic volumes compared to reserved or dedicated infrastructure. Serverless computing has become a default choice for many event-driven backends, APIs with unpredictable traffic patterns, and glue code connecting other cloud services, while sustained high-throughput or latency-critical workloads often still favor traditional or containerized always-on infrastructure.

Key Concepts

• Cloud provider manages server provisioning, scaling, and patching automatically
• Automatic scaling from zero to high concurrency based on real-time demand
• Pay-per-use billing based on actual execution time/resources, not reserved capacity
• No cost during idle periods, since compute scales down to zero
• Most commonly implemented as Function as a Service (FaaS)
• Also encompasses serverless databases, containers, and messaging services
• Introduces cold-start latency and execution time limits as trade-offs
• Can lead to vendor lock-in around provider-specific APIs and event formats

Use Cases

Frequently Asked Questions

Related in Cloud