This blog post is based on a webinar that explored how institutions develop cross-campus models to meet changing research and computing data needs. The panel featured presenters from OSF Institutions members Duke University, Princeton University, and UC Berkeley, as well as NC State University. View the recording here.
Building Collaboration Around a Complex Research Landscape
The contemporary research landscape is generating unprecedented volumes of data, and the demands placed on universities are growing just as quickly.
As Moira Downey, Interim Director of Research Facilitation Service at North Carolina State University Libraries, explained, “Modern research is increasingly technologically complex…Researchers are measuring more things in greater detail, which results in a deluge of data from an increasingly wide range of sources.” At the same time, researchers must navigate rapidly-shifting regulatory expectations.
Libraries have historically played a central role in research data management, but institutions are now seeing what Downey described as “scope creep…We’re starting to inch up to the limits of libraries’ expertise and capacity.” Knowledge gaps around infrastructure, storage, compliance, and custom tools are widening, and faculty are seeking guidance at a scale and level of complexity that existing units may not have been structured to support.
These shared pressures have prompted institutions to develop new, cross-functional models that combine the strengths of libraries, central IT, research computing, and research administration.
North Carolina State University: Humanizing Research IT Support
NC State’s Research Facilitation Service (RFS) grew out of a campuswide recognition that while many services already existed, they were difficult for researchers to find or navigate. Downey described early findings of their gap analysis: “We had a whole wealth of resources…but researchers weren’t aware that they were there.” And even when they were, “the services are spread across multiple campus units who all have varying ways of getting in touch with them.”
To address this, NC State created a cross-functional service jointly supported by the Libraries, Office of Information Technology, and the research office. The team prioritizes hands-on, relationship-based support.
“We’re not meant to be just another ticketing system,” Downey said. “We’re really putting a face on research computing, and data support. We act as a first point of contact, and our preferred method of operating is sitting down with researchers to dig in and learn about their needs in-depth so that we can develop solutions.”
“We're not only connecting researchers to service providers,” she added. “We're helping service providers connect with one another, and break down some of the communication silos that occur on campuses this size.”
RFS links faculty with storage experts, computational specialists, security teams, Institutional Review Board (IRB) staff, and others. This central coordination also helps surface campuswide challenges that individual units may not otherwise see.
UC Berkeley: A Decade of Data Services Through Library–IT Partnership
UC Berkeley’s Research Data Management (RDM) program emerged a decade ago from a campuswide benchmarking study that revealed significant gaps in research data services and limited coordination between existing units. To address this, the Library and Research IT created a shared, jointly staffed program designed to connect researchers with the right expertise at each stage of the data lifecycle.
Erin Foster, Service Lead for the RDM program, explained that the original goal was not to centralize all data services, but to build a coordinating layer: identifying gaps, aligning existing services, and guiding researchers through options that often sit across multiple units.
As Foster noted, the research environment at Berkeley has shifted in recent years, especially as new policy expectations continue to emerge. Working with security, privacy, and compliance partners to interpret policies and ensure that campus systems and tools appropriately support and secure data has become a core function of the program. Over time, Berkeley has streamlined its service areas to focus on six critical domains: data classification and security, data use agreements, data collection tools, general data management, sharing during research, and storage and transfer.
Foster emphasized that the program’s ability to adapt hinges on campus relationships with the security office, IRB, privacy teams, and others whose work increasingly intersects with research data infrastructure.
Princeton University: Designing a Long-Term Data Stewardship Ecosystem
From the start, Princeton’s research data program was built as a cross-campus partnership between the Library, Office of the Dean for Research, and campus IT. Wind Cowles, Associate Dean for Data Research and Teaching, shared that its development was shaped by collaboration, user-driven design, aligning with institutional priorities, and an “ecosystem approach” that avoids siloed services.
Discussions with faculty, research centers, and administrative partners made clear that Princeton needed not just an open data repository, but an integrated set of services capable of handling large, complex datasets while supporting long-term stewardship. This led to two major initiatives:
- Princeton Data Commons, an open data repository designed to support extremely large datasets and emphasize reusability. Cowles explained that the team didn’t want to build another data repository: “We wanted to do something that would be of unique value for Princeton in being able to store and share out research data.”
- TigerData, the university’s next-generation active data storage and management system. Cowles described its mission as providing "not just static storage…but a data storage and management service that is backed by a strong infrastructure,” capable of making even very large datasets FAIR and manageable over time.
These projects are large-scale collaborative efforts—implementing TigerData alone takes about 35 people across multiple teams and campus units. As Cowles emphasized, "It really does take a village."
Panel Discussion: Learning from Experience
Sophia Lafferty-Hess, Senior Research Data Management Consultant at Duke University Libraries and the session's facilitator, guided the speakers through a discussion about the practical realities of building and sustaining these campus partnerships.
Lafferty-Hess opened by sharing that Duke recently launched a Compute and Data Services Alliance in summer 2024—a collaboration between the Libraries, Office of Research, and OIT. With that context in mind, she asked the panelists: What do you wish you had known when you began this work?
Foster reflected on the challenges of maintaining shared vision between partner units, underscoring the importance of continued “shared visioning around the future of the program." While the partnership between Berkeley's Library and Research IT is strong, each unit maintains its own services and priorities. "It's an ongoing challenge of being in this role, to continue to facilitate that: where do we want this to go? How do we want it to develop?"
For Downey, the struggle was articulating value. "When you talk about the service, it makes sense to you," she said. "Then we would talk to a room full of researchers and get crickets or blank stares." NC State found success by working with communication groups to develop concrete use cases and news stories demonstrating how RFS had helped specific research projects.
Cowles noted two interconnected challenges: recognizing that partner units have many priorities beyond the shared work, and understanding the time investment required. "I wish I had known to pay more attention earlier to how the different groups involved in these partnerships are going to have not competing priorities, but additional priorities," she said. "And the amount of time it takes to build these services, partnerships, and infrastructure."
Building Shared Understanding Across Organizational Cultures
The discussion then turned to a challenge shared among all of the panelists: different campus units often use the same terminology to mean different things.
"When I say archive, that means something different and is heard differently from someone in research computing. Or metadata," Cowles described. At Princeton, this led to the creation of a living project glossary that defined key terms in context.
Downey recalled explaining to IT partners that "a couple petabytes of network-attached storage is not a repository." Her team's solution focused on persistent communication through ongoing meetings with IT colleagues and maintaining open channels for questions in both directions.
Foster highlighted the value of dedicated meeting time with partners, sharing that Berkeley established monthly office hours between the RDM program and the campus unit that signs data use agreements, strengthening a partnership that developed organically over time.
Lafferty-Hess noted that Duke took a similar approach early on, holding listening and share-out sessions that created space for staff across units to explain their services in detail, going beyond what a website could convey.
Funding Models and Navigating Change
When the conversation turned to sustainability, the panelists acknowledged ongoing challenges around both funding structures and leadership transitions. Their institutions have taken varying approaches to cost recovery: Princeton charges a nominal ingest fee for extremely large datasets. Berkeley offers base-level computing and storage at no cost, with paid tiers for expanded resources. NC State is exploring chargeback models as the federal facilities and administrative cost landscape shifts, though this creates tension with libraries' traditional model of free services.
Leadership changes require steady communication. NC State formalized expectations through a memorandum of understanding between sponsoring units. At Princeton, Cowles emphasized the value of persistent relationship-building: each time a new senior leader arrives, the team reintroduces their work and its importance.
Looking Ahead
As research infrastructure grows more complex, institutions are recognizing that no single unit can meet all data and computing needs. The programs at NC State, Berkeley, Duke, and Princeton demonstrate that ongoing collaboration is essential to sustainable research support.
This work takes time, persistent communication, and willingness to navigate different organizational cultures and priorities. But the payoff includes stronger data stewardship, smoother research workflows, and institutional capacity that can adapt to emerging challenges. For institutions pursuing similar initiatives, these experiences demonstrate that strong relationships and collaboration are just as critical as infrastructure for successful cross-campus research support.