NFD and My Adventure in the Cloud Native Project Jungle
summary
Node Feature Discovery (NFD) is an add-on that actually reads kernel and CPU interfaces, detects CPU features, power/turbo, NUMA, PCI devices, OS and kernel configs, and publishes node labels so the Kubernetes scheduler and device plugins can pick nodes. Architecture is a per-node worker plus a master that updates the Node object via gRPC between worker and master; discovery runs on a 60‑second loop and supports local plugin hooks (executable books and feature files). Demo shows NFD creating labels that trigger a GPU device‑plugin; mTLS is supported but certificate provisioning is manual.
AI-written summary · about
chapters (20)
- 0:00NFD introduction and journey
- 1:31Speaker background
- 3:01Becoming NFD maintainer
- 4:47Kubernetes resource management
- 6:54Node features explained
- 8:39How NFD discovers features
- 11:11NFD architecture overview
- 12:14Feature sources and labels
- 13:17CPU and hardware detection
- 14:23Power, kernel and OS detection
- 15:50Memory, NUMA and storage detection
- 16:53PCI, network and IO detection
- 18:20Custom feature detection
- 20:33NFD configuration options
- 21:41Deploying NFD in Kubernetes
- 23:30NFD labels for scheduling
- 25:22Live demo of NFD
- 29:03Future plans and roadmap
- 30:54How to contribute
- 31:17Q&A
related talks
33mNFD and My Adventure in the Cloud Native Project Jungle
30mTaming Unbounded Resources with Node Feature Discovery
34mStreamline Your Cluster with Automated Node Labels,Taints, and Annotations
25mK8s Labels Everywhere! Decluttering With Node Profile Discovery.
26mMake Kubernetes Greater for NFV@Arm
27mCloud Native Landscape for Node.js Developers
23mKeynote: Opening Remarks and Networking Journey to CNFs
01h 21mIntro + Deep Dive: Cloud Native Network Function (CNF) Testbed