Midwest Research Computing and Data Consortium
The past decade has seen explosive growth in the utilization of cloud computing to serve a variety of needs — industry analysts estimate that cloud computing experienced >78% growth during 2020 with further double-digit growth projected for 2021-2025. While much of this growth can be attributed to application hosting and (cloud) storage, there is strong demand for cloud cycles and infrastructure hosting (IaaS, PaaS). As institutions of higher education and non-profit, non-academic organizations (e.g. national research laboratories) evaluate their needs, increasingly there is desire for non-commercial cloud computing; examples of research-serving clouds span the nation (San Diego Supercomputer Center’s cloud compute and storage (also including CloudBank), Oak Ridge National Laboratory’s Cades cloud, Aristotle Cloud Federation, Jetstream/Jetstream2, Chameleon Cloud, CloudLab, Penn State (on-prem cloud), University of Michigan’s Yottabyte (Maize) Cloud, and the globe (CERN, Compute Canada, CSCS, NectarCloud, Inria (France), New Zealand’s NESI). How traditional HPC centers approach cloud computing to support increasingly diverse scientific and engineering research and education endeavors is emerging; further, what skill sets CI professionals will need to have to support cloud workflows and data management is of concern.