Building Unified Global Load Balancing for the Edge With k8gb
summary
k8gb is a CNCF Sandbox global service load balancer implemented as a Kubernetes CRD using CRD and GitOps workflows, embedded CoreDNS gossip and ExternalDNS for delegation. The demo runs two local clusters (EU primary, US secondary), fails over when EU replicas scale to zero and recovers back within the DNS TTL (~30s in the demo). It supports nearest-edge routing, weighted spillover and pod-level health checks, runs on Raspberry Pi or cloud, and the demo works end-to-end.
AI-written summary · about
chapters (10)
- 0:00k8gb project intro
- 1:26Global load balancing challenges
- 3:30Traffic control in Kubernetes
- 5:59k8gb architecture
- 9:50Live demo setup
- 13:04Failover between clusters
- 14:28Traffic steering recovery
- 16:14Network partitions handling
- 18:41Project status metrics
- 20:29Q&A
related talks
32mCloud Native Global Load Balancer for Kubernetes
22mCloud Native Multi Cluster/Multicloud Global Load Balancer for Kubernetes
05mProject Lightning Talk: What's New in k8gb: CNCF's Multicluster Global Balancer
40mMulti-Cloud Global Content Distribution at Cloud Native Speeds- 40mMulti-Cloud Global Content Distribution at Cloud Native Speeds
42mSIG Multi-Cluster Deep Dive
33mResilient Multiregion Global Control Planes With Crossplane and K8gb
11mLightning Talk: The Adaptive Edge: Balancing Cost Efficiency, Usability, and Performance Tradeoffs