TCU Cyberinfrastructure Study 2017-2019
Summary Report and Recommendations
The American Indian Higher Education Consortium (AIHEC) conducted a review and documented the state of cyberinfrastructure (CI) readiness and/or implementation at the nation’s Tribal Colleges and Universities (TCUs). The review, conducted from 2017 to 2020 under a grant from the National Science Foundation (NSF award 1655185), involved 34 TCU site visits. The study has provided valuable information about the status of CI at the TCUs and more important, it provided the foundation for the exploration of strategies for building the capacity of TCUs to incorporate CI resources in their research and education programs. This report summarizes the study’s findings and suggests a framework for broader adoption of CI at TCUs and MSIs.
State of the Current TCU CI
While the TCU site visits revealed a wide range of IT capabilities and challenges among the colleges, the following table describes the most common issues and the extent to which they are shared across campuses.
|1||82%||Not enough IT strategic planning|
|2||74%||IT staffing shortages|
|3||71%||Faculty LMS Training|
|5||65%||Technical equipment refresh rates|
|8||53%||IT professional development|
|9||50%||Enterprise Resource Planning systems|
|10||50%||Fiber-optic cabling issues|
Cyberinfrastructures at the TCUs
The TCU CI study concluded that there is significant potential for TCUs to incorporate data science into their STEM education and research programs, even as many TCU IT departments are challenged by fundamental IT operational and infrastructure issues associated with campus networks and systems that must be addressed. While many TCUs appreciate and embrace the importance of engaging CI to promote scholarship and research within their institutions, a comprehensive strategy involving the broader broadening participation in STEM stakeholder community is needed to support their engagement. Given the challenges, the AIHEC TCU CI study team envisions a pathway to CI-readiness through a national CI engagement strategy that would provide a framework through which any TCU (or more broadly, any small, relatively under-resourced higher education institution) can develop strong CI-enabled STEM research and education programs that are responsive to local/regional needs and opportunities regardless of the current status of their CI readiness.
The TCU CI Study revealed a wide range of IT capabilities and challenges among the colleges. Identified CI issues fell into three general categories:
- Resource Challenges: Budgetary limitations cause many TCUs to prioritize other needs over those of their IT departments. This tendency to underprioritize IT department needs contributes to every other technical issue identified by the TCU CI study team. Without adequate budgetary support, the TCU IT departments are unable to develop long-term strategic plans or maintain a technical staff with the combined skills necessary to effectively operate and maintain the colleges’ numerous systems. The IT departments that are understaffed must provide their staff extensive cross-training and skills development that address gaps in credentialed personnel.
- Technical IT Infrastructure issues: Many of the TCUs have significant deficiencies in their basic physical infrastructure. The TCUs’ network challenges include flat (non-routed) campus networks, heavy use of NAT, improper firewall placement and configuration, and fundamental challenges with wireless networks. Campuses with deficiencies in their campus network architecture and wireless systems would benefit from basic engineering assistance in identifying and implementing improvements needed to establish a stable and reliable campus network.
- Internet connectivity: During the time the study was conducted, the average TCU campus Internet connection speed increased from 236Mbps to 449Mbps, due in part to the technical recommendations of this study and technical assistance offered by the TCU CI team. Approximately a third of the TCUs are served by a state or education network and the rest are served by private Internet service providers, and in some instances at excessively high cost.
Programmatic Drivers of CI at TCUs
The AIHEC study team has made some general observations regarding the STEM programmatic demand for CI resources. Although significant differences among the colleges exist, TCUs (and likely MSIs more broadly) can be grouped into three general categories in terms of the current relevance of and demand for CI resources:
- Colleges offering AS, BS and MS degree programs in STEM disciplines, with faculty working on research projects/programs that provide research opportunities for undergraduates. Currently most research being conducted at these institutions is not particularly data intensive;
- Colleges that offer AS and BS degree programs in STEM disciplines, but whose faculty are not actively engaged in research. There are significant opportunities to strengthen and expand these STEM programs through access to CI resources and faculty commitment to development of research programs, particularly through collaborations with colleagues at regional institutions;
- Colleges that offer a limited number of STEM courses that meet general education requirements of non-STEM degree programs but do not offer STEM degree programs. These institutions are likely to have a limited history of engagement with NSF programs.
Colleges at every stage of STEM program development can participate in a TCU-wide effort to make data science and research computing research and education resources available to all faculty and students, with the possibility of access to these resources having a transformative impact on the communities they serve. We recommend an approach that is inclusive of institutions within each category of CI readiness involving strategies based on the framework described below.
A national TCU/MSI CI framework
A coordinated national strategy should be developed involving an array of broadening participation in STEM stakeholders that will support TCUs/MSIs moving their STEM research and education programs to a level appropriate to their institutional goals, the priorities of the communities they serve, and the needs of their students. This national CI framework should include a strong focus both on academic and technical components of TCU/MSI CI readiness with an emphasis on partnerships that provide technical, training and collaborative research opportunities. Critical to a national strategy of CI engagement with TCUs and MSIs generally is an aggressive outreach effort, coordinated and driven by MSIs.
1. IT departments/infrastructure
Human capital. A multi-institutional community of practice (CoP) for MSI IT department staff that provides economies of scale for professional development services, sharing of best and promising practices, and keeps the CoP membership apprised of new and emerging technologies and practices. Professional development opportunities organized for multiple campuses would be a high priority.
Infrastructure upgrades. Building campus technology infrastructure with a focus on providing reliable access (both on campus and remotely) to national data science/research computing resources. This involves both campus technology upgrades and broadband connectivity, particularly to regional or national research and education (R&E) networks. MSIs should have access to reliable and unbiased assistance in equipment configuration and hardware acquisition decision-making.
Remote access to compute resources. Establishing and operating a local HPC facility at most TCUs is not feasible or sustainable. Instead, TCUs should be supported in accessing remote compute resources at low cost, e.g. Open Science Grid and XSEDE.
2. Academic programs
- A community of practice (CoP) in data science will provide a vehicle for identifying skills and knowledge development needs and sharing best practices, particularly in support of the inclusion of data science in the STEM curriculum. It would provide faculty activities that promote and support access and interaction with regional and national CI resources and partners that will strengthen and expand the colleges’ research and education programs.
- A data sciences fellowship program should be developed and offered to faculty at all participating institutions. The fellowship would provide training on data and analytical practices and AI/ML methods of knowledge discovery. It would provide a vehicle for generating opportunities for collaboration with colleagues at other MSIs, mainstream institutions and research centers. Faculty who complete the program would be prepared to teach and/or facilitate data science courses with domain-specific applications, as well as recruit and support students wishing to pursue a data science-related career pathway.
Academic programs offered by TCUs must be enhanced through online courses that fill course and curriculum gaps in a range of technical areas associated with data science and research computing. A data science curriculum with an emphasis on artificial intelligence/machine learning (AI/ML) will be developed and delivered through a combination of online and onsite modalities to students at all MSIs choosing to participate. The curriculum will include courses in computer science, informatics and research computing with specific science domain applications based on existing STEM programs at the institutions. MSIs may choose to adopt and teach the curriculum or support their students access by allowing courses offered remotely by adjunct (non-TCU) faculty to supplement their existing STEM degree programs.
Continual alignment with CI workforce needs. Access to online data science courses would allow TCU programs to be significantly more responsive to evolving workforce demands and opportunities for TCU students. TCU students need access to training and education opportunities that prepare them for CI technical jobs as well as to prepare them to pursue their education goals (undergraduate and terminal degree) that require data science/research computing skills and applications. The colleges (in coordination with the national CI community) should monitor changes in technical workforce requirements and associated career opportunities for TCU students. These changes would drive changes in the curriculum being made available to all TCU students or for adoption by faculty at individual TCUs.
3. Research collaborations
Every TCU/MSI serves a community or region with specific research priorities and needs but has limitations in terms of faculty with research experience and, particularly at primarily teaching institutions, time and support for faculty to develop research programs. CI enables faculty and students to engage in research collaborations that align with community needs and the research interests of the larger research community. Research collaborations among TCU faculty and regional universities should be expedited, and policies put in place at institutions interested in growing their research programs that support and encourage their development. A broad research experiences for undergraduates (REUs) initiative would provide opportunities for students to work on data intensive research projects with researchers at universities and research centers and significantly reinforce the academic and career pathway in data science.
4. Community engagement
Data science resources provided through TCUs to the community (schools, Tribal agencies, organizations) will support broader adoption of data science tools and methods to support economic, health, social and environmental priorities, particularly with the rapidly expanding application of AI/ML to support design and management of health, social and economic initiatives and services. The transformative potential of data science/research computing can best be realized by tribal and other marginalized populations when the work is driven by the communities themselves. TCUs are ideally placed within their communities to serve as a primary source of training and technical support (in collaboration with national CI partners) in the application of data science tools and processes to improve delivery of services across all community sectors. This will also drive a local demand for a data science/CI workforce.
Suggested Stakeholder Roles in Broadening Participation in CI-STEM
The entire community of stakeholders with an interest in broadening participation in CI-STEM should implement the collective impact model (e.g. NSF INCLUDES) establishing networked improvement communities (NICs) focused on facilitating broader access and use of CI resources to address diverse “wicked” problems, particularly within underrepresented communities, such as climate change, energy sustainability and health disparities. Recruiting students to work on these problems (such as climate change) will serve to motivate students to engage in data science while participating meaningfully in addressing a problem with local relevance. Further, it engages domain scientists that will participate in developing career pathways for students. While the details of a coordinated effort would need to be developed by the stakeholders themselves, the following are some possibilities:
TCUs/MSIs: Develop/adapt a core suite of interdisciplinary data science courses available to students in all STEM programs; offer undergraduate courses that emphasize team science with PBL components involving work with data science tools; modify institutional policies to encourage faculty to be more entrepreneurial in pursuing research and research collaborations; ensure strong campus networks/technology that support faculty and student access to off-site compute resources; establish communities of practice/interest in specific data science applications, technical areas;
National CI resources: (e.g. OSG, TACC, XSEDE): Assist with facilitation of research partnerships; continually update and disseminate data science worker skill/knowledge requirements; work with TCUs/MSIs to “platformize” (develop accessible interfaces, collaboratories) customized for specific data science research applications that expedite MSI engagement of research computing tools; assist TCUs/MSIs in the development of professional development trainings for faculty;
Federal agencies with a strong STEM focus: (e.g. NSF, Department of Energy, USDA): in addition to continuation of current programs targeting STEM at MSIs (TCUP, HBCUP) and undergraduate education (e.g. IUSE), support research on potentially transformative applications of data science in communities served by TCUs/MSIs, with emphasis on AI/ML, data management infrastructure requirements; expand CC* program to include stronger emphasis on building the national research computing ecosystem (e.g. OSG) and other initiatives that are democratizing research computing. Most important, incorporate a significantly greater focus on CI in all programs such as INCLUDES that involve a strong collective impact component.