Hardware¶
This section describes the physical hardware used to build the Kubernetes Home Lab.
The goal is to use compact, power-efficient, and affordable machines while keeping a setup close to a real-world bare-metal Kubernetes cluster.
Hardware overview¶
| Node | Role | CPU | RAM | Storage | Network |
|---|---|---|---|---|---|
| cp-1 | Control Plane | Ryzen 5 3500U | 16GB | 512GB SSD | 2.5 GbE |
| worker-1 | Worker Node | Ryzen 5 5500U | 32GB | 500GB SSD | 2.5 GbE |
| worker-2 | Worker Node | Ryzen 5 5500U | 32GB | 500GB SSD | 2.5 GbE |
Control Plane node¶
The control plane is hosted on a dedicated mini-PC.
Characteristics: - Sufficient CPU power for API server, scheduler and controller manager - Moderate memory footprint (16 GB is sufficient for a single control plane) - SSD storage for etcd and system components - Dedicated Ethernet interface
This node is intentionally not used for application workloads.
Worker nodes¶
Worker nodes are more powerful and designed to run application workloads.
Characteristics: - Higher CPU core count for parallel workloads - More memory for pods and caches - Fast local SSD storage - Wired Ethernet connectivity
This allows realistic scheduling, resource limits and scaling experiments.
Networking hardware¶
All nodes are connected using a dedicated 2.5 GbE Ethernet switch.
Design choices: - Wired network only (no Wi-Fi) - Predictable latency and throughput - No network bottlenecks between nodes
Key hardware design decisions¶
- Single control plane (home lab scope)
- Dedicated nodes (no virtualization)
- Wired Ethernet only
- Over-provisioned worker nodes
- Energy-efficient hardware