Demonstrator 1: Smart Connected Factory of the future

 

Demonstrator 1 showcases how the ELASTIC project deploys a secure and scalable IoT Data Fabric in an industrial manufacturing environment. It connects edge, fog, and cloud layers into a unified system that supports predictive maintenance, real-time robotic control, and cross-site data sharing. The goal is to validate latency performance, data trustworthiness, orchestration efficiency, and scalability under real-world conditions.

Why an IoT Data Fabric for Manufacturing

Manufacturing sites generate large volumes of diverse machine and sensor data. Solely cloud-based systems can cause latency and bandwidth issues, while fully decentralized solutions can lead to operational silos. The IoT Data Fabric offers a balanced architecture:

  • It processes time-critical data at the edge.

  • It enables aggregated insights and model updates in the cloud.

  • It supports secure collaboration without compromising data ownership or privacy.

This approach reduces downtime, improves decision-making, and enhances factory-wide visibility.

Core Components

Demonstrator 1 integrates several key technologies to establish a flexible and trustworthy compute layer:

  • WebAssembly (Wasm): Provides portable, sandboxed modules for real-time industrial analytics.

  • Trusted Execution Environments (TEEs): Protect sensitive data and code during execution on untrusted devices.

  • Federated Learning: Trains models locally and shares updates without moving raw data.

  • eBPF and XDP: Improve networking efficiency and reduce latency for control and monitoring.

  • Function-as-a-Service (FaaS): Orchestrates resource-efficient workload scaling and dynamic deployment across layers.

These technologies work together to create a secure, adaptable compute fabric for industrial use.

Security and Trust Mechanisms

Industrial environments demand robust security. In Demonstrator 1, this is achieved through:

  • A zero-trust architecture across edge and cloud components.

  • Remote attestation to verify device integrity before execution.

  • Fine-grained access control policies that govern data movement and user permissions.

  • Secure enclaves (TEEs) to isolate critical operations, reducing the attack surface.

These measures protect both the data and logic that drive industrial processes.

Orchestration Across Edge, Fog, and Cloud

Manufacturing environments are dynamic. Resource availability, latency requirements, and workloads can change quickly. The orchestration framework in Demonstrator 1 manages this by:

  • Placing tasks where they can meet latency and reliability requirements.

  • Scaling services in response to processing load.

  • Shifting non-urgent workloads to energy-efficient tiers during off-peak times.

  • Continuously monitoring performance to adapt workload placement and maintain service-level objectives.

This ensures smooth operations and resilience as conditions evolve.

Validation Scenarios

Predictive Maintenance

Machine data streams are processed at the edge for near-real-time anomaly detection. Wasm modules perform the analysis, and TEEs provide secure execution to protect proprietary data and models.

Cross-Factory Knowledge Exchange

Each factory site trains machine learning models locally and shares parameters via federated learning. This enables model improvement without exposing raw industrial data, aligning with data protection and compliance needs.

Real-Time Robot Control

Robotic control systems require low-latency feedback loops. By running control logic close to factory robots and optimizing network behavior with eBPF, the system achieves consistent response times. TEEs secure command integrity and confidentiality.

Expected Outcomes and Metrics

Demonstrator 1 aims to measure:

  • Improved responsiveness in control systems.

  • Lower bandwidth consumption between factory and cloud.

  • Greater energy efficiency through smart workload distribution.

  • Proven scalability across multiple industrial locations.

  • Enforced end-to-end trust and data security.

Results will contribute to a 6G-ready reference architecture tailored for industrial environments.

Conclusion

Demonstrator 1 illustrates how a unified IoT Data Fabric can deliver secure, real-time, and scalable performance in industrial settings. Combining WebAssembly, TEEs, federated learning, and modern orchestration enables flexible and secure deployment of applications across edge and cloud infrastructure.