5 minutes reading time
(906 words)
Sharing the Load: Why ORION-DBs and Its New FORCE11 Working Group Matter
On 4 September 2026, CWTS Leiden hosted the seminar "Collective action for open research information: introducing ORION-DBs and a new FORCE11 working group." The session brought together Cameron Neylon, Bianca Kramer, Alysson Fernandes Mazoni, Rodrigo Costas, Nees Jan van Eck and Najko Jahn to present a collaboration that has been developing for more than a year, and a new FORCE11 working group intended to give it a longer-term structure.
If you work anywhere near scientometrics or open research information, this is worth paying attention to. And if, like me, you are directly involved in the provisioning of one of these datasets on Google BigQuery, it raises some particularly interesting questions.
What ORION-DBs brings together
ORION-DBs stands for Open Research Information Online Databases. Behind it are teams from MultiObs, Sesame Open Science, SUB Göttingen, and CWTS that had independently been publishing full, queryable copies of open research information resources, including OpenAIRE, OpenAlex and Crossref data, through Google BigQuery.
What changed is that these teams decided to stop working in parallel and seek global-scale coordination.
That matters because many scientometric questions are difficult to answer through APIs alone. Whole-dataset, longitudinal analysis, and especially analysis combining multiple sources, has traditionally required significant local infrastructure or access to proprietary databases. A task reserved only to a handful of players with enough resources to crunch such data.
ORION-DBs makes that kind of work easier by bringing several collections into the same cloud environment. Existing use cases already range from bibliodiversity and research-cost studies to institutional coverage analysis and combinations of ORCID, OpenAIRE Graph, OpenAlex, SciELO, arXiv, and DataCite data.
The point is not simply to host more data. It is to make open research information genuinely usable and cost-effective at scale.
What changed is that these teams decided to stop working in parallel and seek global-scale coordination.
That matters because many scientometric questions are difficult to answer through APIs alone. Whole-dataset, longitudinal analysis, and especially analysis combining multiple sources, has traditionally required significant local infrastructure or access to proprietary databases. A task reserved only to a handful of players with enough resources to crunch such data.
ORION-DBs makes that kind of work easier by bringing several collections into the same cloud environment. Existing use cases already range from bibliodiversity and research-cost studies to institutional coverage analysis and combinations of ORCID, OpenAIRE Graph, OpenAlex, SciELO, arXiv, and DataCite data.
The point is not simply to host more data. It is to make open research information genuinely usable and cost-effective at scale.
From shared datasets to shared infrastructure
The new FORCE11 working group formalises this coordination. Its aims include developing shared documentation and tooling, expanding the network of data providers, collecting user stories and training material, and exploring alternative technological platforms.
That last point is the one I find especially important.
BigQuery works extremely well for this type of analysis. It solves difficult problems around storage, authentication and large-scale querying, without requiring every participating organisation to build its own infrastructure.
But it is still a proprietary commercial platform.
For publicly funded research infrastructures, this creates a tension that is difficult to ignore. If open research information is increasingly used to support research assessment, funding, monitoring, policy and AI-based research intelligence, we also need to think carefully about who controls the infrastructure through which that information is accessed and analysed.
This is not an argument against BigQuery per se. It is an argument for avoiding a situation in which one commercial platform becomes the only viable way of doing this work.
That last point is the one I find especially important.
BigQuery works extremely well for this type of analysis. It solves difficult problems around storage, authentication and large-scale querying, without requiring every participating organisation to build its own infrastructure.
But it is still a proprietary commercial platform.
For publicly funded research infrastructures, this creates a tension that is difficult to ignore. If open research information is increasingly used to support research assessment, funding, monitoring, policy and AI-based research intelligence, we also need to think carefully about who controls the infrastructure through which that information is accessed and analysed.
This is not an argument against BigQuery per se. It is an argument for avoiding a situation in which one commercial platform becomes the only viable way of doing this work.
Where I think the work needs to go next
From my perspective provisioning the OpenAIRE Graph on BigQuery, coordination should not stop at the destination.
We also need more agreement on the path from a raw data snapshot to a usable, reproducible dataset: how schemas are handled, how provenance and reproducibility are preserved, and how datasets are structured so that cross-source querying does not become another bespoke exercise every time.
The working group can help here by encouraging common approaches to processing, documentation and join conventions.
I also think the vendor-lock-in question needs a structural answer, rather than simple backup plans or platform selection strategies.
The criteria already discussed for alternative platforms — publicly governed infrastructure, open-source technology and the ability to analyse data without heavy local requirements — sound remarkably close to what research e-infrastructures are designed to provide.
A hybrid model may therefore be more realistic than a wholesale replacement. Large-scale analysis could remain in BigQuery (or similar) where it is currently the most practical option, while smaller and medium-scale analysis could increasingly move to publicly governed, community-hosted infrastructure, particularly when based exclusively on open research information.
Politically, that matters. Publicly funded research infrastructures should be able to offer credible non-proprietary paths for accessing and analysing the information they steward, even when commercial cloud services remain part of the technical landscape.
The same questions were discussed one week later at STI 2026 in Antwerp, where the ORION-DBs community presented "Sharing the Load: Open Research Information Resources for the STI Community in the Cloud" and demonstrated cross-dataset analysis in practice.
We also need more agreement on the path from a raw data snapshot to a usable, reproducible dataset: how schemas are handled, how provenance and reproducibility are preserved, and how datasets are structured so that cross-source querying does not become another bespoke exercise every time.
The working group can help here by encouraging common approaches to processing, documentation and join conventions.
I also think the vendor-lock-in question needs a structural answer, rather than simple backup plans or platform selection strategies.
The criteria already discussed for alternative platforms — publicly governed infrastructure, open-source technology and the ability to analyse data without heavy local requirements — sound remarkably close to what research e-infrastructures are designed to provide.
A hybrid model may therefore be more realistic than a wholesale replacement. Large-scale analysis could remain in BigQuery (or similar) where it is currently the most practical option, while smaller and medium-scale analysis could increasingly move to publicly governed, community-hosted infrastructure, particularly when based exclusively on open research information.
Politically, that matters. Publicly funded research infrastructures should be able to offer credible non-proprietary paths for accessing and analysing the information they steward, even when commercial cloud services remain part of the technical landscape.
The same questions were discussed one week later at STI 2026 in Antwerp, where the ORION-DBs community presented "Sharing the Load: Open Research Information Resources for the STI Community in the Cloud" and demonstrated cross-dataset analysis in practice.
From Open Research Information to Open Research Intelligence
These discussions also connect closely with OpenAIRE's growing collaboration with the Centre for Science and Technology Studies (CWTS).
The week of the FORCE11 working group announcement, OpenAIRE and CWTS met at CNR in Pisa to explore how open research information can become a stronger foundation for research analytics, monitoring and responsible research assessment.
The workshop looked at practical issues including coverage, affiliations, citations, deduplication, classifications, provenance and validation, but also at a broader question: how do we move from Open Research Information to Open Research Intelligence?
For me, the connection is clear. OpenAIRE brings community-governed research information infrastructure and the OpenAIRE Graph; CWTS brings internationally recognised expertise in bibliometrics, research analytics and responsible assessment.
As AI and research-intelligence agents become more influential, keeping the evidence, methods and assumptions behind them transparent will become even more important.
ORION-DBs, the FORCE11 working group and the OpenAIRE–CWTS collaboration all point in the same direction: open research information should not only be accessible. The infrastructure and methods used to turn it into knowledge should also remain open, transparent and responsibly governed.
The week of the FORCE11 working group announcement, OpenAIRE and CWTS met at CNR in Pisa to explore how open research information can become a stronger foundation for research analytics, monitoring and responsible research assessment.
The workshop looked at practical issues including coverage, affiliations, citations, deduplication, classifications, provenance and validation, but also at a broader question: how do we move from Open Research Information to Open Research Intelligence?
For me, the connection is clear. OpenAIRE brings community-governed research information infrastructure and the OpenAIRE Graph; CWTS brings internationally recognised expertise in bibliometrics, research analytics and responsible assessment.
As AI and research-intelligence agents become more influential, keeping the evidence, methods and assumptions behind them transparent will become even more important.
ORION-DBs, the FORCE11 working group and the OpenAIRE–CWTS collaboration all point in the same direction: open research information should not only be accessible. The infrastructure and methods used to turn it into knowledge should also remain open, transparent and responsibly governed.
Stay Informed
When you subscribe to the blog, we will send you an e-mail when there are new updates on the site so you wouldn't miss them.
