Skip to main content
Definition Website Computer and Technology

Serverless Computing

Subject: Computer and Technology noun /ˈsɜːr.vər.ləs kəmˈpjuː.tɪŋ/ 1 definition

  1. Definition (noun)

    Entry #79 · Updated

    Serverless Computing is a cloud computing execution model in which a cloud provider dynamically provisions, scales, and maintains the underlying servers, operating systems, and runtime environments required to execute application workloads. Despite the name, physical and virtual machines still run the code; the serverless boundary lies in abstracting host management, capacity planning, and infrastructure patching away from developers. Workloads execute on demand—often scaling down to zero capacity when idle and expanding automatically during traffic bursts—so organizations pay strictly for active execution duration and consumed resources rather than pre-allocated servers.

    Common Forms

    • Function as a Service (FaaS): Executes short-lived, stateless code functions inside ephemeral sandboxes in response to discrete events—such as HTTP requests, message-queue arrivals, or storage uploads—automatically scaling concurrency per invocation.

    • Backend as a Service (BaaS): Supplies turnkey, API-driven backend capabilities—including user authentication, object storage, and push notifications—allowing web and mobile applications to operate without custom server processes.

    • Serverless Containers: Runs packaged OCI container images on managed runtimes that handle request routing, TLS termination, and scale-to-zero orchestration without requiring operators to provision virtual machines or cluster nodes.

    • Serverless Databases: Decouples compute from durable storage so database engines automatically scale processing capacity with query load, pause during inactivity, and meter billing by active compute units and stored data.

    Example

    When a customer uploads a 12 MB product photo to an e-commerce store's cloud object bucket, a storage event triggers a Python image-resizing function on AWS Lambda that writes compressed WebP thumbnails in under 400 milliseconds. During quiet overnight hours with zero uploads, no compute instances remain active or incur charges, whereas a flash sale triggering 3,000 concurrent uploads automatically provisions parallel sandboxes without manual capacity changes.