Edge Computing: Processing Data Closer to Users
Edge computing moves computation and data processing from centralised cloud data centres closer to where data is generated or consumed — at the "edge" of the network. This reduces latency for time-sensitive applications and reduces the bandwidth required to transmit raw data to the cloud for processing.
Types of Edge Computing
- CDN edge compute: Code running on CDN edge nodes (Cloudflare Workers, Fastly Compute@Edge, Lambda@Edge) — executes in the PoP closest to the user. Suitable for: request routing, A/B testing, authentication, personalisation, and API gateway logic. Latency: <10ms in most locations.
- IoT/OT edge: Processing on devices or local servers at the physical location where sensors and operational technology are located — before sending relevant data to cloud. Reduces bandwidth, enables operation when cloud connectivity is unreliable.
- Telecommunications edge (MEC): Compute hosted by mobile network operators at the network edge — ultra-low latency for applications like AR/VR, autonomous vehicles, and industrial automation.
Use Cases for CDN Edge Compute
- Serving personalised content without routing requests to the origin
- JWT validation and authentication at the edge — reducing load on application servers
- Geolocation-based routing and content customisation
- Rate limiting and bot protection at the edge
- A/B test bucket assignment without client-side JavaScript
When We Use Edge Compute
Cloudflare Workers are part of our standard architecture for applications where latency and scale matter. Edge logic should be simple and stateless — complex business logic belongs in the application tier.