Hosting Models
Serverless
A model where you deploy code and the provider runs it on demand, billing only for actual execution with no server to manage or keep running.
Serverless doesn't mean there are no servers. It means you never think about them. You deploy a function or a small app, the platform runs it only when a request comes in, and you pay for the milliseconds it actually executes. Idle costs nothing.
The appeal is real for spiky or unpredictable traffic. A function that runs a thousand times a day costs almost nothing, and one that suddenly runs a million times scales without you touching a config. Cloudflare Workers, AWS Lambda, and the function layers inside Vercel and Netlify all work this way.
The footguns are cold starts, execution time limits, and pricing that gets murky at scale. Long-running or memory-heavy jobs fit poorly, and a runaway function can generate a genuinely alarming bill. It's a scalpel, not a Swiss Army knife.
Go deeper
- Where to Host a Node.js API in 2026PaaS, VPS, or serverless for an Express/NestJS/Fastify backend: what each actually costs, where each breaks, and the memory ceiling nobody warns you about.