Cloud Adoption Survey in Major Facilities and Mid Scale Institutes

Authors: Bruce Berriman, Bob Flynn, Rajiv Mayani, Benedikt Riedel, Mats Rynge, Tyson Swetnam, Amanda Tan, & Karan Vahi

From the report:

This survey targeted personnel from NSF-funded Major Facilities (MFs) and Mid-scale Research Infrastructure (MRIs) awardees to assess their current and future use of cloud services (commercial and academic) and identify barriers to adoption. These facilities handle vast amounts of data (i.e., Petabyte or PByte+ scale), requiring significant storage, processing, and serving resources.

The survey was sent to 288 personnel across 21 MFs and 10 MRIs, yielding responses from 13 facilities, 7 mid-scales and 2 NSF projects. Most respondents identified as technical leaders or managers, with the majority indicating some level of cloud adoption. Organizations reported using major commercial cloud providers (e.g., Amazon, Google, Microsoft), academic providers (e.g., Jetstream), and self-hosted solutions (e.g., OpenStack). Respondents generally recognized cloud benefits, including improved reliability, scalability, on-demand access, and better storage services.

Despite acknowledging cloud benefits, respondents cited major barriers to adoption, including lack of funding, complexity in re-architecting deployments, and workforce skill gaps. Cost concerns - particularly for storage, compute, and data egress - were the top challenge, often leading organizations to adopt hybrid (i.e., on-premise and cloud) strategies. Additionally, many on-prem data centers were found to undercharge for computing, often lacking proper cost models. These findings align with long-standing discussions within the CI Compass Cloud TWG.