A major focus in computing research across industry, academia, and government is to advance the frontiers of computing by experimenting with leading-edge compute platforms.  This was initially at odds with early instances of cloud/warehouse computing that focused on deploying commodity rather than state-of-the-art systems.  However, cloud computing platforms have advanced significantly during the last two decades to the point where there are now many notable examples of their use in research.  Further, the cloud’s “pay as you go” model has proved attractive for research related to machine learning and data analytics that benefit from elastic provisioning of resources. It is noteworthy that NSF has extended its CloudBank portal with a CloudBank Catalog with links to commercial cloud services that can be paid from NSF grants. Hence, increased use of the cloud for computing research could offer new revenue opportunities to cloud providers, in addition to joint research across industry and academia. There will also likely be a crossover from the use of cloud computing for research to its use in teaching, thereby making future generations of computing professionals more accustomed to cloud platforms.

The purpose of this roundtable was to discuss best practices on using the cloud for computing research and the resulting synergistic opportunities across industry, academia, and government.  Some of the discussion topics included:

• Use of cloud computing to broaden the set of institutions that can gain access to large-scale resources for computing research
• Fostering research collaborations across industry, academia, and government (also enabled by data sharing and open-source software)
• Preparing students to be consumers of cloud computing in their future careers
• Enabling computing researchers to be creators and early adopters of leading-edge cloud technologies, while looking under the hood and customizing the systems being used
• Demystifying and capping costs incurred when using the cloud for research.