What is Serverless?
A cloud execution model where code runs in stateless containers spawned per request, billed by execution time rather than uptime.
Serverless is a cloud model where you upload code (typically a function) and the platform handles all infrastructure — container provisioning, scaling, OS patching. You're billed per invocation and execution duration, not per server uptime. Servers still exist; they're just abstracted away. AWS Lambda, Vercel Serverless Functions, Cloudflare Workers, Google Cloud Functions, and Azure Functions are the major implementations. Pros: zero-config scaling (handle 1 RPS or 10,000 RPS without code changes), pay-per-use pricing (great for spiky traffic, expensive for sustained high traffic), no infrastructure management. Cons: cold starts (first request after idle adds latency), execution time limits (typically 10-900 seconds), no persistent state between requests (must use external storage), vendor lock-in is real. Edge Runtime (Vercel/Cloudflare Workers) is a lighter-weight subset with sub-50ms cold starts but more constraints.