A federated policy-governed architecture for sovereign European data spaces

<ns5:p>European data spaces aim to enable trusted, sovereign and interoperable data sharing across organisational boundaries. Existing reference frameworks address parts of the problem in isolation: GAIA-X concentrates on federated trust and self-descriptions, the International Data Spaces Reference Architecture Model (IDS-RAM) emphasises usage control and the connector pattern, the Data Spaces Support Centre (DSSC) Blueprint consolidates governance vocabulary and Simpl-Open provides a vendor-neutral middleware substrate. The integration of semantic interoperability, policy-driven sovereignty, federated computation across the cloud–edge–high-performance computing (HPC) continuum and decentralised governance within a single deployable architecture remains an open design problem. This article proposes and analyses a conceptual, policy-governed federated architecture derived from the architectural work of the Horizon Europe NOUS project (Grant Agreement No. 101135927). The architecture is organised in five functional layers — data, services, compute, federation and governance — and revolves around five interacting roles: the Data Space Participant (DSP), the Data Space Infrastructure Participant (DSIP), the Data Space Common Services (DSCS) and the companion NOUS Data and Infrastructure Agents (NDA, NIA). A clear separation between Control Plane and Data Plane mediates policy negotiation, contract enforcement, federated learning and infrastructure provisioning across cloud, edge, HPC and quantum back-ends. We illustrate the model through two representative scenarios derived from NOUS use cases — crisis management with heterogeneous geospatial and field data, and cross-organisational federated analytics over heterogeneous compute resources — and we map architectural functions onto major European reference initiatives. The contribution is conceptual: the article does not claim empirical validation, performance benchmarks or formal certification of standards compliance. Limitations and a research agenda — prototyping, federated interoperability testing and formal verification of policy enforcement — are stated explicitly.</ns5:p>