Serverless is a cloud execution model where providers manage infrastructure. You write functions, the provider handles scaling. You pay per execution, not for idle time. AWS Lambda, Google Cloud Functions, Azure Functions, Vercel Functions, and Cloudflare Workers are all serverless platforms. You upload code. The platform runs it when needed.
If a function needs to handle 1 request per second most of the time but 1000 requests per second during a traffic spike, the platform automatically scales. You pay only for the execution time. This is different from renting servers where you pay monthly regardless of load. Serverless is ideal for event-driven workloads.
Processing file uploads, sending notifications, API endpoints with variable traffic. You don't waste resources on servers sitting idle. The tradeoff is that serverless functions are stateless and have execution limits. A function might timeout after 15 minutes. You can't store data in memory between executions. You can't run background processes forever.
For these use cases, you need traditional servers or containers. But for many applications, serverless is simpler and cheaper than managing your own infrastructure. Serverless democratizes deployment. You don't need to understand scaling, load balancing, or infrastructure. You just deploy code.