Kubernetes Networking Panel: Collaborate on the Spec, Compete on Execution
summary
Kubernetes enforces per-pod L3 IPs so any implementation just has to deliver IP reachability; vendors compete on rendering while SIG keeps the intent minimal. Implementations shown include CoreOS’s Flannel (VXLAN tunneling used in many installs), Calico (IPv6-native routing), Contiv/others integrating with underlay SDN and ACI for VM+container apps. Policy uses label-driven NetworkPolicy plus RBAC and namespaces; about half the audience already uses network policies, and multi‑tenancy and traffic shaping are next priorities.
AI-written summary · about
chapters (24)
- 0:00Intro
- 1:57Scope
- 4:07L3 Networking
- 5:31L4 Networking
- 7:49Hook vs Hook
- 12:38Labels
- 15:27Different Models
- 17:20RoleBased Access
- 19:23Security
- 20:22Multitenancy
- 22:07Softwaredefined Networking
- 24:29IP Networks
- 26:20Networking Model
- 27:09Use Cases
- 28:59SDN
- 29:54Hybrid Use Cases
- 30:36Automation
- 31:27IPv6 Migration
- 33:29Security Issues
- 34:14L2 vs L3
- 35:28Choice
- 38:17Whats the next big thing
- 39:06Open issues
- 39:40Work arounds
related talks
37mPanel Discussion: Cloud Native Networking State of the Union
35mPanel: Kubernetes Networking: Present and Future
32mThe Easy--Don't Drive Yourself Crazy--Way to Kubernetes Networking
21mFrom Squire To Network Knight: Defending the Kubernetes Castle the Cloud Native...
38mKubernetes Networking Infrastructure Offload
34mLife of a Packet [I]
40mCloud, Kubernetes, and Service Networking - Taming the Turtles
19mSIG-network: Updates and Future Directions