

K2D by Portainer
May 17, 2024
Neil Cresswell and Steven Kang from Portainer discuss K2D, a project merging Kubernetes tooling with Docker to manage containers on edge devices. They delve into the architecture, motivations, and limitations of K2D, highlighting its efficiency and compatibility with small devices like Raspberry Pis. The podcast explores running Kubernetes tools on Docker without a full Kubernetes distribution and the future possibilities of lightweight container configurations.
AI Snips
Chapters
Transcript
Episode notes
K2D Enables Kubernetes API on Docker
- K2D runs a partial Kubernetes API on top of Docker Engine without a full Kubernetes control plane.
- It allows Kubernetes tools on tiny devices like old Raspberry Pis with far less resource use than K3S.
K2D Ultra Light Resource Footprint
- K2D requires only about 20 MB RAM when idle, burning almost no CPU without Kubernetes API requests.
- This efficiency makes it ideal for devices with as little as 512 MB RAM where K3S needs much more.
K2D Emulates Single-Node Cluster
- K2D emulates a single-node Kubernetes cluster on one Docker host without clustering.
- It translates Kubernetes API calls to Docker commands, with no swarm support yet.