Building to last - Strategic Approaches to a sustainably funded Institutional Research Data Management

At research institutions, an array of infrastructures for institutional RDM is emerging which includes both highly specialized applications and standard services.[3] They are developed in response to specific needs, and their implementation fosters not only technical expertise but also valuable knowledge about processes.[7] These efforts often rely on local resources or third-party funding.[6] However, when such resources are depleted or discontinued, these services may be left behind as "zombies" or shut down. In these early phases, those operating such services often face legal risks they may not be aware. Not every initiative, tool or service deserves to be institutionalized. It would be more beneficial for the scientific community and host institutions if users—or even paying customers—could determine the value of start-up-like initiatives, rather than leaving it to bureaucratic chance. Effective operational models for sustainable RDM rely on transparent divisions of labor, legal certainty, and clearly defined responsibilities.[4] Proven models lead to more efficient resource use—both within individual services and across the landscape of academically used services. Transparent accounting of resources is essential for a legally sustainable operational model. Conceptual platforms are required[2] that could serve as hosts for operational model templates, which include initiatives like the NFDI,[5] state initiatives or, as a subsidiary option, central IT services at universities.[8] Research institutions must actively shape their RDM strategies and integrate the diverse landscape of local, regional, and (inter-)national services into a coherent overall offering.[3][9] Both the bottom-up-driven NFDI and the European Open Science Cloud, as well as state-level initiatives, play an increasingly important role in this. There are multiple examples of bottom-up initiatives showing potential for integration into a distributed service landscape. Popular services with longer development histories, such as Galaxy, OMERO, InvenioRDM, or Chemotion, coexist alongside approaches like DataPLANT's GitLab-based DataHUB.[3] Service operations span from foundational infrastructure like servers and storage, ideally provided by computing centers, to research-proximate activities closely aligned with domain-specific communities.[8][10] Funding models must account for these different actors and should aim for a balance of fairness, simplicity, and practicality. When designing them, a mix of low-threshold access for general use and effort-based billing for high-demand users should be considered (i.e., Basic Service, Paid Extra).[9] This necessitates service governance and coordination mechanisms—currently being tested through NFDI—that bring together stakeholders from various domains and align their activities to create synergies. To build trust in the long-term operation of services, their entire lifecycle—from conception and development to operation, scaling, and eventual decommissioning—should be actively supported. Sustainable, cooperative and federated RDM requires not only mutual trust but also external service integration, the ability to extend one's own services to third parties, and more flexible staffing models. The considerations laid out in this paper go back to discussions in the "Arbeitskreis Forschungsdatenmanagement" (AK FDM)[3] embedded in the permanent state-wide exchange on how to foster a vibrant landscape of research data management initiatives.