GOAL 3: RESEARCH COMPUTING
Goal 3 (Research Computing) provides all faculty and graduate students with expanded access to research software, data management and computing services and upgrades the campus infrastructure to provide networking speeds and high performance computing capabilities required by faculty in data intensive research fields.
- 3.1Upgrade network capacity to provide high-speed connectivity on the campuses and
to regional and national research networks.
- 3.2 Determine the research computing capabilities required by each school and determine
the best structures to enhance capabilities.
- 3.3 Create a research computing investment strategy to support a growing research portfolio, facilitate successful faculty recruitment and retention, and enable more faculty to use computing resources in their research and scholarship.
Goal 3: Provide an Advanced Cyber Infrastructure to SupportGreater Research Activity and to Enable Faculty in More Disciplines to Incorporate Computing Capabilities in Their Research and Creative Works.
The research activities and aspirations of the faculty require more capable and easily accessed research and scholarly computing capabilities. Multiple schools' research programs are becoming more data intensive and research computing is increasingly relevant to scholarship in many non-science disciplines. Most faculty require access to similar baseline research computing capabilities that would be inefficient for each to maintain on their own. Research growth plans and competitiveness to recruit and retain the best faculty with research computing intensive projects will depend, in part, on offering more capable computing infrastructure with enhanced features to secure, analyze, and transport increasingly larger data sets. Further, offering all faculty members and many students access to research computing support will promote non-sponsored scholarship, help newer faculty develop their research programs, and support teaching in academic programs, such as data sciences.
Today, the campus network does not offer fast enough speeds for researchers to efficiently transfer very large data sets securely or to access the research infrastructure at a partner institution without degraded performance that could impact research results. Storage and data management services are insufficient to handlelarge data sets and data requiring stringent security protections. Many of the elements important to research (e.g., high performance computing, GIS, are organized within separate departments and are not easily discovered or accessed by faculty in anotherpart of the University. Finally, the high performance research computing infrastructure is aging and without dedicated funding to support its renewal.
This focus area of the plan will meet these challengesby defining each school's research computing needs and determining a strategy for each to develop enhanced capabilities. Additionally, continued improvements will be made to the foundational networking, security and software capabilities important to research and scholarship.
1. Upgrade network capacity to provide high-speed connectivity on the campuses and from the University to regional and national research networks.
The initiative creates significantly higher speed connectivity to Internet 2 and invests in campus networking infrastructure to provide researchers with connection speeds of 10 gig, and lays the technical foundation to provide 100 gig in the future. These changes create twenty times faster network speeds and offer much greater capacity to move large data sets, access remote computing facilities and work with collaborators around the world. Implementing a science DMZ on the campus network will provide researchers with a portion of the network with security settings and equipment configurations optimized for access to high performance computing resources and the bulk transfer of large amounts of data to and from collaborators. The DMZ structure will enable researchers to work within security structures appropriate to the data they are handling without degrading network performance and hampering the conduct of the research. Partnerships will be pursued with NYSERNET and the other SUNY University Centers to bring greater network bandwidth and redundancy to Binghamton, while also improving high speed connectivity to research infrastructure within NY and nationally.
Advising canbe supported with a more holistic picture of a student's academic and co-curricular engagement by consolidating the information in disparate systems. Further, by integrating data from these source systems, students and appropriate faculty and staff can be given dashboards and alerts to monitor if they are on track to meet academic goals, graduation requirements, their engagement with the Fleishman Center or their participation in internships and other co-curricular activities.
2. Determine the research computing capabilities required by each school and select the best structures to enhance capabilities.
Researchers in multiple disciplines and transdisciplinary centers require research infrastructure and software to handle large amounts of data, compute at a high speed or high throughput, store data at differential levels of security and access speed. Further, the frequency and duration at which the infrastructure is needed varies greatly by research program. As each School articulates their research computing needs, the University will be able to identify areas of common need that could be addressed collaboratively and areas of more distinct needs. ITS will continue to provide foundational network, security and access management, and software licensing to facilitate both collaborative and distributed research computing services and needs.
Strengthening the shared infrastructure will enable the University to adopt any number of models to support its future research computing needs. One option is a highly-distributed model in which each school arranges for their own compute, storage and support infrastructure. Or, consideration could be given to establishing a University-wide research computing service to offer more broadly needed services for which there are economies of scale and a scarcity of expertise such as high performance computing, research data management, research data storage, specialized research software, and research application development. Or, the University may opt for a hybrid model that combines school specific capabilities, capabilities resident in one school but provided to the whole university,and some shared services offered by a research computing group. Shared web sites and cross trained staff supporting research can be developed to make it easier for faculty to discover the full breadth of available support and quickly access the right expertise. As the model develops, a research computing coordinating committee of faculty and IT providers should be created to assist with the coordination of services and the identification of new needs.
Longer term, on-campus capabilitiesto support research will be extended with capabilities offered by other institutions and organizations. Pre-negotiated agreements will be developed to facilitate direct access to cloud providers of computing platforms with predictable costs and contract terms. Future improvements to networking and virtualized infrastructure will also provide flexibility to order compute and storage capability on-line and have it provisioned on-demand via on premise or cloud resources. Advanced research infrastructure will be facilitatedby improved connectivity to the other SUNY University Centers and NYSERNET. Networking partnerships will provide faculty access to private cloud research computing facilities and collaborators at other institutions.
3. Create a research computing investment strategy to support a growing research portfolio, facilitate successful faculty recruitment and retention, and enable more faculty to use computing resources in their research and scholarship.
Creating and sustaining better research computing capabilities will require investment and a financial strategy to secure funding for on-going operations and regular replacement of infrastructure as it ages. Today, no individual organization is funded to offer a complete set of research services, nor is there adequate capacity to offer baseline services (e.g., storage, access to computing cycles, software) to all faculty. This strategic initiative will develop a long-range financial strategy for research computing to enable requisite expansion to support faculty research interests and provide sustainable infrastructure. Through a consultative process, effective methods will be defined to share costs of baseline and advanced research computing services.
Consideration will be given to effective practices at other institutions. In concept, research computing capabilities function like a core research facility and similarly require multiple sources of funding. As result, other research universities approach the funding challenge in multiple ways. University operating budgets and seed funds are used to create new capacity and fund baseline services to faculty without external funding. Overhead recoveries enable investment to expand capabilities and services in areas that align with areas of increasing research interest and sponsored funding. Finally, direct charges to grants help fund operations of some services and offer faculty an avenue to buy services and avoid using grant funds to create separate computing infrastructure. A comprehensive business plan and funding proposal tailored to Binghamton will be developed as part of the implementation of this strategy.