Towards FAIRness using an Interoperable Research Data Architecture Developed within NFDI-MatWerk

In an era where data-driven research is reshaping scientific discovery, ensuring FAIR data is paramount. Building on these principles, we present a shared service architecture conceptualized within the German NFDI-MatWerk. The architecture, initially developed based on the requirements and needs of the materials science and engineering (MSE) community, can be extended and adapted for different domains, for national (e.g., NFDI) and international (e.g., EOSC and Gaia-X) scientific communities. We add a layer of interoperability to the existing infrastructure components, allowing their wide range spanning diverse research institutions and computational centers to be connected, regardless of where they are stored. Crucially, each layer of the architecture also systematically addresses various facets of the FAIR principles - from data ingestion and metadata management to data representation exemplified through the use of FAIR Digital Objects (FDOs), significantly accelerating progress towards data FAIRness. Demonstrating the real-world impact of this approach, the proposed architecture has been applied in selected NFDI-MatWerk Infrastructure Use Cases (IUCs), thereby confirming its alignment with research data management best practices in MSE research contexts. Each IUC uses a tailored combination of services, represented through workflows, acting as bridges between the architecture and practical research needs. These workflow representations make the architecture more accessible, illustrating how components can be connected and applied, serving as blueprints adaptable by researchers, supporting their journey towards transparent, reproducible, and collaborative research. This architecture is therefore a core element in the broad effort to make research data FAIR through its interoperable, adaptable design and practical implementation.