AquaINFRA and the EOSC Federation: Charting a Course for Sustainable Open Science

On November 3, 2025, representatives from 13 research organisations and national initiatives gathered in Brussels for the EOSC Symposium to sign a Memorandum of Understanding that formally launched the European Open Science Cloud Federation. Among the signatories was Kimmo Koski, Managing Director of CSC -- IT Center for Science in Finland, signing on behalf of the Finnish national EOSC Node. For AquaINFRA, that signature marked more than a milestone in European science policy. It opened the door to what may become the project's most important legacy: a sustainable home for the research infrastructure we have spent three years building.

A New Chapter for EOSC

When AquaINFRA began in January 2023, the European Open Science Cloud was still largely a vision. EOSC-Core services existed -- authentication, a resource catalogue, a marketplace -- but there was no operational infrastructure where research projects could host their services and expect them to persist beyond individual funding cycles.

That picture has changed dramatically. In October 2024, the European Commission launched the EOSC EU Node, the first operational node of the federation, providing researchers across Europe with compute resources, a tools hub, and federated authentication through a single entry point. By the time of the Brussels Symposium a year later, the federation had grown to include 13 candidate nodes spanning national infrastructure providers (Finland, Germany, Italy, Poland, Slovakia, and the Netherlands through SURF), thematic research communities (CERN for physics, BBMRI-ERIC for biomedical research, Life Sciences Connect for life sciences, PaNOSC for photon and neutron science), and cross-cutting data and computing services (EUDAT and the Digital Twin of the Ocean).

The EOSC EU Node had already welcomed 4,000 users and distributed over 2.5 million virtual credits in its first year of operation. The federation was no longer a concept document. It was becoming infrastructure.

What This Means for AquaINFRA

AquaINFRA has built an operational research platform for marine and freshwater science. The Data Discovery and Access Service (DDAS), developed by the National Land Survey of Finland and hosted at CSC, connects over 15 major European data sources -- from the European Data Portal and Copernicus Marine Service data to PANGAEA, the CUAHSI Water Data Center, and Finnish Environment Institute monitoring records -- through a single search interface. The Virtual Research Environment (VRE), built on the Galaxy platform and hosted at the University of Freiburg, enables cloud-based reproducible analysis. The AquaINFRA Interaction Platform (AIP), developed by 52North, ties discovery and analysis together.

These are working services. Researchers have used them to detect a 30-year decline in water clarity in the Gulf of Riga, model harmful algal bloom risk in the Catalan Sea, and assess freshwater biodiversity patterns across 378 European river catchments. Published Galaxy workflows have been successfully transferred between Baltic Sea sub-basins, demonstrating that methods developed in one region can be reused in another.

But working services need somewhere to live after December 2026, when AquaINFRA's Horizon Europe funding ends. The Galaxy instance needs compute resources -- virtual machines, CPU, RAM, and storage. The DDAS needs server infrastructure and maintenance. The AIP needs hosting. Open-source code can be copied and modified freely, but running a live platform that researchers depend on requires sustained commitment.

This is where the EOSC Federation becomes relevant; not as an abstract policy framework, but as a practical answer to a concrete question: who hosts the infrastructure after the project ends?

The Nice Brokerage Event: Turning Strategy into Action

In late January 2026, members of the AquaINFRA consortium travelled to Nice for the EOSC Winter School Brokerage Event, organised by the EOSC Association with support from the EOSC Gravity project. The event was designed for precisely the kind of matchmaking AquaINFRA needed: connecting research projects with EOSC nodes that could host their services.

The team held bilateral meetings with four potential node partners:

The Finnish EOSC Node (CSC) is the most natural fit for the DDAS, which is already hosted at CSC. The National Land Survey of Finland is working to formally onboard the data discovery service through the Finnish node, which entered its build-up phase in March 2025 and offers services including fairdata.fi, the LUMI supercomputer, and cloud and storage infrastructure. The motivation is clear: integration with EOSC would give the DDAS visibility under the EOSC umbrella and connect it with the federation's capabilities for monitoring, authentication, and catalogue discovery.

The EUDAT Node provides another existing connection. AquaINFRA datasets are already being published through EUDAT's B2SHARE service with persistent identifiers, making them discoverable through OpenAIRE and the broader EOSC resource catalogue. This ensures that even if active platform services face hosting challenges, the data itself -- with its DOIs and standardised metadata -- will remain accessible.

The German EOSC Node (NFDI) was discussed with several German AquaINFRA partners, including DKRZ, IGB, and 52North. The NFDI4Earth initiative already runs a Galaxy instance, which could provide a natural integration point for AquaINFRA's VRE workflows and tools.

The Data Terra Node, a French research infrastructure with thematic hubs covering atmosphere, solid earth, ocean, land surfaces, and biodiversity, offered thematic alignment, particularly for AquaINFRA's marine and environmental data services.

The consortium also met with the EOSC Beyond project to discuss integrating AquaINFRA's pygeoAPI services with EOSC monitoring capabilities.

A Multi-Node Strategy

What emerged from these discussions is not a single-node solution but a federated approach that mirrors the EOSC Federation's own architecture. Different components of AquaINFRA align with different nodes:

• The DDAS, with its Finnish development team and existing CSC hosting, fits naturally within the Finnish EOSC Node

• Published datasets, with their persistent identifiers and standardised metadata, are already preserved through EUDAT

• VRE tools and Galaxy workflows could potentially be sustained through the German node's NFDI4Earth Galaxy infrastructure

• Marine data connections align with the Digital Twin of the Ocean node

This multi-node approach reflects a broader reality about domain-specific research infrastructure: platforms that bridge disciplines -- as AquaINFRA bridges freshwater and marine science -- do not fit neatly within a single thematic or national node. The EOSC Federation's governance will need to accommodate services that span multiple nodes, and AquaINFRA's experience is helping to test how that works in practice.

The Funding Question

Honesty demands acknowledging what the EOSC Federation does not yet solve. All 13 first-wave candidate nodes, with the exception of the EOSC EU Node itself (which is funded by the European Commission), participate on an in-kind basis without dedicated funding. The second-wave enrolment call, which closed on 18 February 2026, similarly carries no budget. This means that nodes must absorb the cost of hosting additional services from their existing institutional resources.

For a service like the DDAS, where the hosting institution (CSC) is already committed to the Finnish EOSC node and has the infrastructure in place, this is manageable. For compute-intensive services like the VRE, which requires ongoing access to virtual machines and processing capacity, the sustainability question is more open.

The EOSC Financial Sustainability Task Force has produced 15 recommendations for sustaining the federation beyond the current Horizon Europe programme period ending in 2027. A key principle in those recommendations is inclusiveness: whatever funding model is adopted, costs should not act as a barrier to participation. But recommendations are not yet commitments, and the post-2027 governance model remains under discussion.

Looking Ahead

AquaINFRA's final year is not just about completing deliverables. It is about ensuring that the infrastructure we have built -- and the scientific capabilities it enables -- outlasts the project that created it. The EOSC Federation provides the most credible pathway for that, but it requires active engagement, bilateral agreements with nodes, and realistic planning about what can be sustained and how.

The work that began with the EOSC Federation's formal launch in Brussels and continued with the brokerage meetings in Nice will intensify through 2026. The target is to finalise integration arrangements by autumn 2026, ahead of the project's December endpoint. Not everything may be sustained in its current form, but if the data remains discoverable through EOSC, the workflows remain published and reproducible on Galaxy, and the core discovery and access services find a home within the federation, then the investment that the European Commission and 21 partner institutions have made in AquaINFRA will continue to generate returns for European water science long after the grant agreement closes.

The EOSC Federation is still young; formally launched only months ago and still building its operational capabilities. AquaINFRA is one of the first Horizon Europe research infrastructure projects to navigate the transition from project funding to federation hosting in real time. The lessons learned will matter not only for aquatic research, but for every domain-specific infrastructure project that follows.