The major cloud providers, including Amazon Web Services, Google Cloud and Microsoft Azure, are now participating in the U.S. National Science Foundation’s flagship research program on big data, Critical Techniques, Technologies and Methodologies for Advancing Foundations and Applications of Big Data Sciences and Engineering (BIGDATA). Amazon, Google and Microsoft will provide cloud credits/resources to qualifying NSF-funded projects, enabling researchers to obtain access to state-of-the-art cloud resources.
The BIGDATA program involves multiple directorates at NSF, as well as the Office of Financial Research, and anticipates funding up to $26.5 million, subject to availability of funds, in Fiscal Year 2017. Additionally, AWS, Google and Microsoft will provide up to $9 million (up to $3 million each) in the form of cloud credits/resources for projects funded through this solicitation.
This novel collaboration combines NSF’s experience in developing and managing successful large, diverse research portfolios with the cloud providers’ proven track records in state-of-the-art, on-demand, cloud computing. It also builds upon the shared interests of NSF and the cloud providers to accelerate progress in research and innovation in big data and data science—pivotal areas that are expected to result in tremendous growth for the U.S. economy.
The BIGDATA program encourages experimentation with real datasets; demonstration of the scalability of approaches; and development of evaluation plans that include evaluation of scalability and performance among competing methods on benchmark datasets—all of which will require significant storage, compute, and networking resources, which can be provided by the cloud vendors through their participation.
Proposals requesting cloud credits/resources must adhere to a 70:30 split between NSF funding and cloud resources, respectively, and must not request less than $100,000 for cloud requests. Thus, if a project requests $700,000 in NSF funds, then it may request a maximum of $300,000 in cloud credits/resources from one of AWS, Google, or Microsoft, or a minimum of $100,000. This minimum budget requirement underscores key objectives of the BIGDATA program, which include supporting experimentation with data and studying data scaling issues.
Proposal submissions are due March 15 through March 22, 2017 (and no later than 5 p.m. submitter’s local time on March 22). All those interested in submitting a proposal to the BIGDATA program should refer to the solicitation for details. All proposals that meet NSF requirements will be reviewed through NSF’s merit review process. For proposals that request cloud resources, reviewers will additionally be asked to evaluate: (1) the appropriateness of the requested use; (2) whether the specific use of cloud resources has been adequately justified through an annual usage plan; and (3) the estimate of the amount of resources needed and the corresponding resource request budget (in dollars). The requests for cloud resources should not only include resources required for the experimentation phase, but also for usage over the duration of the project (e.g., software development and testing and code debugging).