Serverless Computing
Definition
Running code without managing servers yourself - the cloud provider handles all the server management. Like ordering takeout instead of cooking at home.
Use Cases
- Coca-Cola: Processing data from connected vending machines (IoT) to support monitoring and operations workflows. — Used AWS Lambda with AWS IoT services to run event-driven processing when device messages arrived, avoiding always-on servers for bursty device traffic. (Reduced operational overhead by offloading server management and improved responsiveness by processing events as they arrived.)
- Bustle Digital Group: Handling traffic spikes and background processing for digital media workloads. — Adopted AWS Lambda for event-driven tasks and scaling during unpredictable traffic patterns, integrating with other AWS managed services. (Improved ability to scale for peak demand while reducing the need to provision and manage server capacity ahead of time.)
Provider Equivalents
- AWS: AWS Lambda
- Azure: Azure Functions
- GCP: Google Cloud Functions
- OCI: Oracle Functions
Frequently Asked Questions
- What's the difference between Serverless Computing and containers (like Kubernetes)?
- Serverless runs your code on-demand and the cloud provider manages scaling and servers for you. Containers package your app and dependencies, but you (or a platform like Kubernetes) still manage capacity, scaling rules, and cluster operations. Serverless is often simpler for event-driven tasks; containers are often better for long-running services and more control over runtime and networking.
- When should I use Serverless Computing?
- Use it for event-driven or spiky workloads where you want minimal operations work: APIs and webhooks, file/image processing after uploads, scheduled jobs, queue/message processing, and lightweight data transformations. It’s especially useful when traffic is unpredictable or you want to pay only when code runs.
- How much does Serverless Computing cost?
- Costs are typically based on number of invocations (requests), execution duration, and allocated resources (like memory/CPU). You may also pay for related services such as API gateways, queues, databases, logging/monitoring, and network egress. It can be very cost-effective for intermittent workloads, but always-on high-throughput workloads may be cheaper on reserved containers/VMs depending on usage patterns.
Category: cloud
Difficulty: intermediate
Related Terms
See Also