Edge Deployment
The Ona Platform runs fully on-premise on Ubuntu 22.04 — no AWS, no internet connection required after initial setup. All services run as Docker containers on a single machine, with PostgreSQL as the data plane, a local filesystem volume for ML artifacts, and Redis for caching.
This section covers everything needed to deploy, operate, and build against an edge installation.
What You Need
| Component | Minimum | Recommended |
|---|---|---|
| CPU | 4 cores | 8+ cores |
| RAM | 8 GB | 16 GB |
| Disk | 20 GB free | 50+ GB |
| GPU | None | NVIDIA 4+ GB VRAM |
| OS | Ubuntu 22.04 LTS | Ubuntu 22.04 LTS |
8 GB RAM supports inference only. 16 GB supports site-specific training alongside all running services.
Pages in This Section
| Page | What it covers |
|---|---|
| Deploy | Hardware preflight, install script, post-install verification, accessing the UI |
| Architecture | Service map, data planes, training pipeline, FastAPI wrapper pattern |
| Training CLI | ona_train.py command reference with examples |
| Security | Auth model, network isolation, ISO 27001 behavioral test suite |
| SDK Edge Mode | ONA_MODE=edge, endpoint env vars, available clients |