Just returning home after attending an @NSF workshop on the future of cloud research infrastructure in Princeton where I had the privilege, along with @rabernat, of representing the @pangeo_data project.

https://sites.google.com/view/workshop-on-cloud-cri

Some quick take aways: 1/5

The @NSF_CISE community sees cloud as a test bed for studying system architecture. Focus is mainly on system design and performance. Public clouds don’t offer sufficient flexibility to test new hardware configurations so they will likely keep building smaller private clouds. 2/5

The domain sciences, especially those with big data (e.g. @pangeo_data), are looking to @NSF for a vision for supporting research applications on the cloud. Public clouds seem to offer the necessary components, scalability, and ease of use to service these communities. 3/5

Should @NSF_CISE be the org that provides the domain sciences with access/funding to the cloud? Maybe not. It seems unlikely that the same community that is interested in developing cloud test beds has interest in providing infrastructure for domain science applications. 4/5

I’m hopeful that #NSFfunded orgs like #CloudBank can provide the right interface between the public cloud providers and domain sciences. https://www.nsf.gov/news/news_summ.jsp?cntn_id=298986&org=CISE