D5.4 Connectors for Common European Data Spaces

This deliverable presents the initial strategy and technical foundation for integrating the LUMI AI Factory (LUMI AIF) with Common European Data Spaces through standardized connector components. These connectors are essential for enabling secure, policy-compliant, and interoperable data exchange between LUMI AIF and external data ecosystems, in alignment with the European Data Strategy. The report outlines the architectural principles, component selection criteria, and deployment roadmap for connectors that support data sovereignty, trust, and interoperability. It emphasizes the role of connectors as operational enablers of governance logic, capable of enforcing usage policies, managing identity, and facilitating dynamic contract negotiation across heterogeneous systems. Three key connector technologies are evaluated:

• REMS, for entitlement workflows and access governance in sensitive domains like health and genomics.

• Eclipse Data Space Connector (EDC), for standardized, sovereign data exchange across research and language data spaces.

• Simpl middleware, for scalable integration with multiple European data spaces and support for semantic interoperability and policy enforcement.

The report identifies high-value data spaces for initial integration, including the European Open Science Cloud (EOSC), European Language Data Space (LDS), and the European Health Data Space (EHDS). It also highlights the strategic relevance of REMS in EHDS and its operational use in the Genome Data Infrastructure. A phased deployment plan is proposed, starting with sandbox environments and MVP connectors for EOSC and LDS, followed by regulated integrations using Secure Processing Environments and REMS. The plan includes technical baselines, governance mechanisms, and operational models to ensure compliance with evolving European standards such as IDS-RAM, Gaia-X, DSSC blueprints, and the Data Act. Looking forward, the report recommends expanding connector coverage, enhancing compliance tooling, and aligning with strategic initiatives like Exa4Mind and the EuroHPC Federation Platform. These efforts will position LUMI AIF as a federated, trusted node in the European AI and HPC ecosystem, supporting scalable, responsible, and innovation-driven data use.