FaaS
Definition
Function as a Service - execute individual code functions in the cloud without server management, enabling scalable and efficient computing.
Use Cases
- Netflix: Automated security and operations tasks (e.g., responding to events, enforcing policies, and running small automation jobs). — Netflix has publicly described using AWS Lambda as part of its AWS-based tooling to run event-driven automation and operational workflows, triggered by cloud events and integrated with other AWS services. (Faster automation of routine tasks and improved operational efficiency by running code on-demand without provisioning or managing servers for each small job.)
- The Coca-Cola Company: Backend processing for connected vending/IoT and digital experiences, handling event-driven workloads. — Microsoft has published case studies describing Coca-Cola using Azure services, including Azure Functions, to process events and integrate systems in a serverless, scalable way. (Improved scalability for bursty event traffic and reduced operational overhead by relying on managed, on-demand compute for discrete tasks.)
- The New York Times: Media processing pipelines (e.g., transforming and processing images and other assets as part of publishing workflows). — The New York Times has publicly discussed building cloud-based processing pipelines on AWS; serverless functions (commonly AWS Lambda in such architectures) are often used to trigger transformations from storage events and orchestrate steps with managed services. (More elastic processing for large volumes of assets and faster turnaround for publishing workflows without maintaining dedicated processing servers.)
Provider Equivalents
- AWS: AWS Lambda
- Azure: Azure Functions
- GCP: Cloud Functions
- OCI: Oracle Functions
Frequently Asked Questions
- What’s the difference between FaaS and PaaS?
- FaaS runs small, event-triggered functions and automatically scales them per request; you deploy individual functions, not a full app server. PaaS (like app hosting platforms) runs long-lived applications or services where you deploy an app and the platform manages the underlying infrastructure, but you still think in terms of an always-on web app/process.
- When should I use FaaS (serverless functions)?
- Use FaaS when your work can be broken into small tasks triggered by events (file uploads, HTTP requests, queue messages, database changes, scheduled jobs). It’s especially useful for spiky or unpredictable traffic, lightweight APIs, automation, ETL steps, and background processing. Avoid it when you need long-running processes, very low-latency warm execution at all times, or highly specialized runtime control.
- How much does FaaS cost?
- Most FaaS pricing is based on (1) number of invocations/requests, (2) execution duration, and (3) memory/CPU allocated (often tied together). You may also pay for related services like API gateways, message queues, logging/monitoring, and outbound network traffic. Costs are usually low for intermittent workloads, but can grow with high request volume, long execution times, chatty logging, or heavy network egress.
Category: cloud
Difficulty: advanced
Related Terms
See Also