Behind the Scenes of Building a Gateway API Dataplane...
Talkat KubeCon + CloudNativeCon + OpenInfra Summit + PyTorch Conference China 2026 · 2026-09-07..09-09·
summary
Compares two Envoy-based Gateway API implementations and gives concrete trade-offs: default deployment uses one Envoy fleet per gateway in controller land space for isolation and security; cluster modeling chooses sharing vs per-route clusters with a merge backend option to collapse 50 duplicate clusters into one; config delivery shows State-of-the-World xDS rebuilds cost (1,000 routes ≈ 600 ms) and an endpoint fast path plus a 10s TTL to avoid 503s; extensibility options include CRDs, Envoy patching and a proposed MuxCache for decoupled incremental updates.
AI-written summary · about
chapters (5)
- 0:00Gateway API dataplane intro
- 2:08What is Gateway API
- 6:02Envoy Gateway design
- 18:54Airlock Microgateway
- 30:31Key takeaways & Q&A
related talks
29mTaming the Traffic: Select...
27mEnvoy Gateway: An in-Depth Guide of Its Benefits, Use Ca...
20mEnvoy Gateway Project Update
33mEnvoy Gateway Project Update
39mEnvoy Gateway: The API Gateway in the Cloud Native Era
14mEnabling Multi-Cluster Load Balancing with Gateway API and Envoy Gateway
15mFrom ingress-nginx to Envoy Gateway at Zapier: The...
24mBuild Your Own Envoy Control Plane
