TwinEdge Edge Overview
TwinEdge Edge is the on-premise component of the TwinEdge platform. It runs on your local network, connects directly to industrial equipment, and provides local dashboards, ML inference, and data buffering.
Architecture
┌─────────────────────────────────────────────────────────────────────┐
│ TwinEdge Edge │
├─────────────────────────────────────────────────────────────────────┤
│ │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌────────────┐ │
│ │ OPC UA │ │ ML Engine │ │ Storage │ │ Alert │ │
│ │ Server │ │ (ONNX) │ │ (SQLite) │ │ Engine │ │
│ └──────┬──────┘ └──────┬──────┘ └──────┬──────┘ └─────┬── ────┘ │
│ │ │ │ │ │
│ └────────────────┴────────────────┴────────────────┘ │
│ │ │
│ ┌─────────┴─────────┐ │
│ │ Node-RED │ │
│ │ Flow Engine │ │
│ └─────────┬─────────┘ │
│ │ │
│ ┌─────────┴─────────┐ │
│ │ Local Dashboard │ │
│ │ (Port 8080) │ │
│ └───────────────────┘ │
│ │
└────────────────────────────────┬────────────────────────────────────┘
│
│ HTTPS / MQTT
│
▼
┌───────────────────────┐
│ TwinEdge Cloud │
│ (api.twinedgeai.com) │
└───────────────────────┘
Key Features
Protocol Hub
TwinEdge Edge acts as a protocol translator and data aggregator:
- OPC UA Server: Exposes all connected devices through a unified namespace
- Multi-Protocol: Connect Modbus, EtherNet/IP, S7, and MQTT devices
- Tag Browser: Discover and map device tags automatically
Local ML Inference
Run machine learning models directly on the edge:
- ONNX Runtime: Execute models trained in the cloud
- Low Latency: Sub-100ms inference for real-time decisions
- Model Updates: OTA model deployment from cloud
Local Storage
Buffer data locally for reliability:
- SQLite Database: Store telemetry during network outages
- Parquet Export: Export historical data for analysis
- Configurable Retention: Set data retention policies
Alert Engine
Respond to anomalies instantly:
- Local Alerts: No cloud dependency for critical alerts
- Multi-Channel: LED indicators, local buzzers, relay outputs
- Cloud Sync: Forward alerts to cloud when connected
Local Dashboard
Monitor equipment without internet:
- PWA: Install as app on tablets
- Real-time Charts: Live data visualization
- Touch Optimized: Designed for industrial HMI use
Deployment Modes
| Mode | Description | Use Case |
|---|---|---|
| Development | Demo data, simulators enabled | Testing, demos |
| Production | Real devices only, no test data | Live deployments |
| Offline | No cloud sync, fully autonomous | Air-gapped sites |
Hardware Requirements
Minimum (Raspberry Pi 4)
- CPU: 4 cores @ 1.5GHz
- RAM: 2GB
- Storage: 16GB SD card
- Network: Ethernet recommended
Recommended
- CPU: 4+ cores @ 2GHz
- RAM: 4GB+
- Storage: 32GB+ SSD
- Network: Gigabit Ethernet
Installation
Option 1: Pre-Built Image (Recommended)
- Download the TwinEdge Edge image from downloads.twinedgeai.com
- Flash to SD card using Balena Etcher
- Boot your Raspberry Pi
- Access setup wizard at
http://twinedge.local:8080
Option 2: Docker Compose
# Clone the repository
git clone https://github.com/twinedge/twinedge-edge.git
cd twinedge-edge
# Start services
docker-compose up -d
# Access dashboard
open http://localhost:8080
Option 3: Manual Installation
See Installation Guide for detailed steps.
Getting Started
- Connect to Network: Plug in Ethernet or configure WiFi
- Access Dashboard: Navigate to
http://twinedge.local:8080 - Link to Cloud: Enter your organization token
- Add Devices: Configure industrial protocol connections
- Start Monitoring: View real-time data on the local dashboard
Next Steps
- Protocols & Connections - Connect industrial devices
- Local Dashboard - Configure the HMI
- OTA Updates - Manage firmware updates
- Troubleshooting - Common issues and solutions