Serverless
A cloud model where you deploy code or use a database without managing servers: the platform scales automatically and charges only for what runs, including zero when there is no usage.
Serverless does not mean "no servers": it means the servers stop being your problem. You ship a function, an API, or use a managed database; the provider handles provisioning machines, scaling up when traffic rises, and releasing resources when it drops. Billing follows actual usage: per request, per execution time, or per data read and written, instead of a machine running (and billed) 24/7.
The most common forms are:
- Functions (FaaS): pieces of code that run in response to events: an HTTP request, a webhook, an uploaded file. Examples: AWS Lambda, Cloudflare Workers, Vercel Functions.
- Serverless databases: databases that scale and bill by usage, such as Supabase, Convex, and Turso, removing manual instance sizing.
The strengths: cost proportional to use (great for irregular traffic or early projects), zero system maintenance, and automatic scaling. The trade-offs: less control over the environment, per-function execution time limits, and the *cold start* (the small extra latency when an idle function has to "wake up").
Concrete example: a contact form receiving 30 submissions a day. On serverless, each submission runs a function for a few milliseconds and the monthly bill is cents; on a VPS, the same task would require a server running all month.