EdgeFlow is designed to run on a wide variety of hardware, from tiny single-board computers to industrial gateways and cloud infrastructure. This guide helps you choose the right platform for your needs.
Raspberry Pi
Raspberry Pi is the primary development and recommended platform for EdgeFlow. These affordable, well-documented boards offer excellent community support and are perfect for home automation, prototyping, and small-scale deployments.
| Model | Status | RAM | CPU | Best For |
|---|---|---|---|---|
| Pi 5 | ✅ Full Support | 4-8 GB | Quad-core 2.4GHz | AI/ML workloads, complex automations, multi-protocol hubs |
| Pi 4 | ✅ Full Support | 2-8 GB | Quad-core 1.5GHz | Most users - excellent balance of performance and cost |
| Pi 3 B+ | ✅ Full Support | 1 GB | Quad-core 1.4GHz | Basic flows, simple sensor monitoring, budget setups |
| Pi Zero 2 W | ✅ Full Support | 512 MB | Quad-core 1GHz | Compact deployments, remote sensors, space-constrained installs |
| Pi Zero W | ⚠️ Limited | 512 MB | Single-core 1GHz | Very basic flows only - no AI features, limited node count |
💡 Recommendation
For most users, we recommend the Raspberry Pi 4 (4GB) as the best starting point. It offers excellent performance for typical home automation tasks at a reasonable price point. If you plan to run AI/ML models locally or need maximum performance, consider the Pi 5 (8GB).
Other Single Board Computers
EdgeFlow supports various ARM and x86-based single-board computers. While Raspberry Pi receives primary development focus, these alternatives offer unique advantages for specific use cases.
🍊 Orange Pi
Tested
Officially tested on Orange Pi 5. Offers similar capabilities to Raspberry Pi with often better availability. The Orange Pi 5 features an RK3588S chip with NPU for AI acceleration.
- Orange Pi 5 - Full support
- Orange Pi 4 - Community tested
- Orange Pi Zero - Limited support
🦴 BeagleBone
Community
Popular in industrial and educational settings. BeagleBone boards feature real-time PRU processors, making them excellent for precise timing applications and industrial I/O.
- BeagleBone Black - Community supported
- BeagleBone AI-64 - For AI workloads
- PocketBeagle - Compact option
🟢 NVIDIA Jetson
Recommended for AI
The best choice for AI/ML-intensive edge deployments. Jetson boards feature dedicated CUDA cores and Tensor cores for accelerated inference, perfect for computer vision and advanced analytics.
- Jetson Orin Nano - Entry-level AI
- Jetson Orin NX - Professional AI
- Jetson AGX Orin - Maximum performance
🪨 Rock Pi
Community
Radxa Rock boards offer powerful Rockchip processors with good Linux support. Some models include NPU for AI acceleration at competitive prices.
- Rock 5B - High performance
- Rock 4 SE - Budget friendly
- Rock Pi S - Ultra compact
Industrial Hardware
For production deployments in factories, warehouses, and commercial environments, industrial-grade hardware offers enhanced reliability, extended temperature ranges, and professional support contracts.
| Manufacturer | Product Line | Key Features | Use Case |
|---|---|---|---|
| Advantech | UNO Series, MIC Series | Fanless, wide temp range (-20°C to 60°C), DIN rail mount | Factory automation, SCADA integration |
| Siemens | SIMATIC IoT Gateways | OT/IT integration, industrial protocols, MindSphere ready | Manufacturing, process control |
| Dell | Edge Gateway 3000/5000 | Rugged design, cellular connectivity, remote management | Retail, logistics, remote monitoring |
| Moxa | UC Series | Serial connectivity, industrial certifications | Legacy system integration |
| OnLogic | Helix, Karbon Series | Fanless, shock/vibration resistant | Transportation, outdoor deployments |
⚠️ Industrial Deployment Note
Industrial deployments often require additional considerations: power redundancy, network segmentation, compliance certifications (UL, CE, ATEX), and integration with existing SCADA/DCS systems. Contact our team for guidance on production deployments.
Cloud & Virtual Environments
EdgeFlow can run in virtualized and cloud environments for development, testing, or hybrid edge-cloud architectures. This flexibility enables centralized management and seamless scaling.
🐳 Docker
Run EdgeFlow in isolated containers for easy deployment, version management, and portability. Ideal for development environments and microservices architectures.
docker run -d \
--name edgeflow \
-p 1880:1880 \
-v edgeflow_data:/data \
edgeflow/edgeflow:latest☸️ Kubernetes
Deploy EdgeFlow as pods in Kubernetes clusters for high availability, auto-scaling, and centralized orchestration. Perfect for managing multiple edge instances from a central control plane.
- Helm charts available
- K3s support for edge Kubernetes
- KubeEdge integration
☁️ AWS
Deploy on Amazon Web Services for cloud-based processing or hybrid architectures. Graviton instances offer ARM-native performance at lower cost.
- EC2 (x86 or Graviton ARM)
- ECS / Fargate containers
- AWS IoT Greengrass integration
☁️ Google Cloud
Run on Google Cloud Platform for serverless deployments and integration with Google's AI/ML services.
- Cloud Run (serverless containers)
- Compute Engine VMs
- GKE for Kubernetes
- Anthos for hybrid edge
Minimum System Requirements
| Component | Minimum | Recommended | AI/ML Workloads |
|---|---|---|---|
| CPU | Single-core 1GHz ARM/x86 | Quad-core 1.5GHz+ | Quad-core 2GHz+ or GPU/NPU |
| RAM | 512 MB | 2 GB | 4-8 GB |
| Storage | 4 GB | 16 GB | 32 GB+ (for models) |
| Network | Ethernet or WiFi | Gigabit Ethernet | Gigabit Ethernet |
| OS | Linux (Debian/Ubuntu recommended), macOS, Windows (Docker) | ||
Choosing the Right Platform
🏠 Home Automation
Recommended: Raspberry Pi 4 (4GB)
Affordable, quiet, low power consumption. Perfect for controlling lights, sensors, and basic automations.
🤖 AI/ML Projects
Recommended: NVIDIA Jetson or Raspberry Pi 5
Hardware acceleration for inference. Run object detection, voice recognition, and predictive models at the edge.
🏭 Industrial/Production
Recommended: Advantech or Siemens gateways
Reliability, certifications, and professional support for mission-critical deployments.
👨💻 Development/Testing
Recommended: Docker on your workstation
Quick iteration, easy reset, consistent environment across team members.
📍 Remote/Compact
Recommended: Raspberry Pi Zero 2 W
Tiny form factor, WiFi built-in, low power. Great for distributed sensor networks.
☁️ Cloud/Hybrid
Recommended: Kubernetes + Cloud provider
Scalability, centralized management, integration with cloud services.