WEDA Architecture
WEDA implements a two-tier architecture—WEDA Cloud and WEDA Edge—where each layer has well-defined responsibilities and communicates through standardized protocols.
Developer Focus Areas
WEDA handles all infrastructure concerns—device connectivity, data synchronization, container orchestration, security, and lifecycle management—so developers can focus on two areas:
1. Frontend Development with WEDA Core APIs
Build custom management interfaces and dashboards using WEDA's RESTful APIs:
- Management Consoles: Web-based interfaces for device monitoring and control
- Data Visualization Dashboards: Analytics dashboards for time-series data and KPIs
- Custom Workflows: Domain-specific operational workflows and automation
Use the Opensource UI Boilerplate to accelerate development with pre-built components.
2. AI Solution Development
Develop and deploy intelligent edge applications:
- AI Model Training: Train custom ML/AI models for specific use cases
- Sub Node Development: Build lightweight applications for constrained devices and sensors
WEDA manages model deployment, versioning, and inference infrastructure. Developers focus on model accuracy and business outcomes.
WEDA Cloud Layer
The cloud layer provides centralized management, orchestration, and development capabilities through API-centric services.
WEDA Core
The core platform exposes RESTful APIs for programmatic access to all platform capabilities:
| Capability | Description |
|---|---|
| Device Management | Centralized device provisioning, configuration, and lifecycle management |
| Data Management | Time-series data collection, storage, and analytics |
| AI Model Management | Model registry, versioning, and deployment orchestration |
| Container Management | Container registry, stack definitions, and deployment workflows |
Enablers
Foundational libraries, SDKs, and utilities providing common functionality for application development and system integration.
WEDA Edge Layer
The edge layer runs on physical devices, enabling local processing, real-time decision-making, and autonomous operation.
WEDA Node
The core edge runtime agent:
| Capability | Description |
|---|---|
| Device Management Agent | Local device control, telemetry collection, and bidirectional communication with WEDA Cloud |
| Data Agent | Local data buffering, preprocessing, and intelligent data forwarding |
| Remote Management | Secure remote access, diagnostics, and OTA updates |
| Digital Twin | Local digital twin representation for edge autonomy |
Container & AI Model Management
- Container Management: Docker-compatible container orchestration
- AI Model Management: On-device model loading, inference execution, and model switching
Protocols & Communication
- Protocols Module: Edge protocol implementations for device connectivity (Modbus, OPC UA, MQTT, etc.)
- WEDA SubNode (Open Source): Lightweight agent for constrained devices and sensors. Developers implement sensor integration logic; WEDA SubNode handles Digital Twin protocol communication with WEDA Core.
Development Resources
- Ready-to-Dev Containers: Pre-configured development containers with toolchains and hardware-specific dependencies (NVIDIA+Ubuntu, NXP+Yocto, Qualcomm, etc.)
- Device Library (Open Source): High-level Python/C# library that abstracts Advantech hardware peripherals—GPIO, sensors, cameras, I/O modules.
Enablers
Edge-optimized libraries providing common edge computing functionality such as data filtering and edge analytics.
Third-Party Edge Applications
Support for deploying third-party edge applications, microservices, and custom business logic alongside WEDA platform services.
Architecture Characteristics
Cloud-Edge Collaboration
- Bidirectional Synchronization: Seamless data and configuration sync between WEDA Cloud and WEDA Edge
- Offline Capability: WEDA Node maintains local autonomy and digital twin state when disconnected from WEDA Cloud
- Intelligent Workload Distribution: Dynamic workload placement based on latency, bandwidth, and compute constraints
Scalability & Extensibility
- Horizontal Scaling: Thousands of edge devices managed from a single WEDA Core instance
- Modular Adoption: Use only the components you need—WEDA Node alone, SubNode only, or the full stack