The Federation Architecture Pattern (FAP) on IoT and AI shows how data in federated ecosystems can be collected from many devices, shared securely, brought together in one place, and then analyzed with Artificial Intelligence to be clearly presented on a dashboard.
Purpose & Value
The goal of the IoT & AI FAP is to make it easy to connect any IoT devices with different data systems and AI tools in a standard and flexible way. It can also run safely inside a company’s own network — even without internet. The FAP creates a trusted setup that lets organizations securely collect, share, combine, and analyze IoT data across different locations and partners.
Multi-Channeling
Gathers edge data and manages multiple data channels.
Standardizing
Standardizes data transfer between sender and receiver using the dataspace protocol.
Aggregating
Enables cloud-based data aggregation.
Using AI
Supports AI-driven data analysis.
Visualizing
Visualizes results as dashboard widgets for each data channel.
Scope & Boundaries
Included in the IoT& AI FAP:
- Unified data pipeline from sensor to dashboard in federated ecosystems.
- Full abstraction of data gathering, data transfer, data aggregation, data analysis, and data visualization.
- Integration of widgets and AI via ORCE.
- Integration with ICAM for secure asset access.
- Usage of trust anchor und policy engine for data flow security.
Excluded from the Iot & AI FAP:
- Dataspace management
- Complex data analytics
- Shopfloor protocols
- Deep integration
This FAP builds as follows:
Feature-FAPs:
- IoT Data Collection
- Data Lake Management
- AI and Visualization
Micro-FAPs (examples):
- Dataspace Connector
FAP Components:
- IoT Domain
- Local management of sensor data
- Data Lake
- Consumption of IoT data streams and data persistence
- AI
- Analytics of data and identification of thresholds
- Dashboard
- Visual representation per data channel.
XFSC Services:
- CAT (Catalogue) – local data container for used services
- ORCE – orchestration of FAP service and simulation IoT backend
- AAS – authentication and authorization service
- OCM/PCM – for organization and participant credential managers
- TSA – specify policies for data connector and user actions.
IoT-AI FAP adheres to
- W3C DID/VC: Decentralized identifiers and verifiable credentials for asset provenance.
- OIDC4VC: Standardized flows for credential issuance.
- DIDComm v2: Secure communication between participants.
- Gaia-X Trust Framework: Compliance, trust anchor, catalogue integration.
- JSON-LD: Linked data for standardized service metadata.
- OpenAPI / GraphQL: Service discovery and integration APIs.
- GDPR: Data minimization and lawful processing of metadata
- Dataspace Protocol
IoT-AI FAP is designed to be used as follows:
Cross-domain Reuse:
- IoT Management Platform/ Shopfloor
- Node-RED (https://nodered.org/)
- Data Space Ready Data Connectors
- https://dssc.eu/
- https://github.com/International-Data-Spaces-Association/ids-specification
- https://projects.eclipse.org/projects/technology.edc
- https://simpl-programme.ec.europa.eu/
- Data Lake Service for data collection and aggregation
- https://cloud.ionos.com/solutions/big-data
- Apache Spark, Kafka, Trino, HDFS, Superset
- AI-supported Analytics
- https://cloud.ionos.de/managed/ai-model-hub
- Sovereign AI
- Business Intelligence (BI) Dashboard
- ORCE (Node-RED)
- https://github.com/eclipse-xfsc/orchestration-engine
Reusable Modules:
- Data Connector
- AI Analytics
- Dashboard UI
Variants:
- Asset Administration Shell
- Multiple Data Space Implementations