NSF Major Facilities Cloud Use Cases and Considerations

Published January 12, 2024.

From the "NSF Major Facilities Cloud Use Cases and Considerations" publication:

The National Science Foundation (NSF) supports over 25 Major Facilities (MFs) that serve as cornerstones for the science community. These MFs, characterized by their continuous operations, large-scale and sophisticated data collection, and broad user communities, represent long-term investments intended for multi-decade operations.

Cloud computing presents a wealth of possibilities for these MFs. It offers an extensive range of services, including data storage, archival, processing, and sophisticated data access. Scalability is a distinct advantage of cloud platforms, allowing facilities to transition smoothly from initial development stages to high-demand phases. This shift, however, does not have to be all-or-nothing. MFs can leverage the cloud selectively, optimizing their resources without resorting to an all-in approach.

Case studies, such as those of the National Center for Atmospheric Research (NCAR), IceCube Neutrino Observatory (IceCube), and the combined Seismological Facility for the Advancement of Geoscience (SAGE) and Geodetic Facility for the advancement of Geoscience (GAGE) facilities, operated by the EarthScope Consortium, offer valuable insights. They underline the benefits of the cloud while also highlighting challenges, particularly in comparison to traditional on-premises (on-prem) infrastructure.

When considering cloud adoption, cost emerges as a pivotal factor. Costs are incurred for data storage, egress, compute resources, serverless functions, and application programming interface (API) requests. It is critical for MFs to understand these costs and factor them into operating budgets and planning.

A significant challenge that MFs might encounter is "vendor lock-in." Transitioning between vendors can become costly, primarily because of differing APIs, tools, and services offered by various cloud providers. MFs adopting cloud tools should employ strategies to mitigate vendor lock-in.

As the cloud becomes more integral to operations, training assumes paramount importance. A workforce adept at navigating cloud technologies will be crucial for MFs to harness the full potential of these platforms.

In essence, cloud computing offers a transformative opportunity for NSF MFs. While it introduces myriad complexities, its advantages are manifold. By viewing cloud adoption as a flexible, case-by-case decision rather than an absolute “all-or-nothing” option, MFs can make informed choices about the level and extent of cloud adoption, which has the potential to significantly amplify their capabilities, enhance research quality, and ensure efficient resource utilization. This report encourages MFs to educate themselves on the fundamentals of cloud computing in order to better understand and leverage the cloud tools and computing model where they best fit MF’s specific needs, optimizing benefits while effectively managing challenges and complexities.

This report was authored by: Bruce G. Berriman, Brian Dobbins, Jeremy Fischer, Bob Flynn, Jeffrey Glatstein, Rajiv Mayani, Loïc Pottier, Craig Risien, Benedikt Riedel, Mats Rynge, Erik Scott, Tyson Swetnam, Amanda Tan, Chad Traband, Karan Vahi, Don Brower, and Charles Vardeman

View Publication