Exploring Clouds for Acceleration of Science
What is the E-CAS Project?
In 2018, with support from Amazon Web Services (AWS) and Google Cloud Platform (GCP), Internet2 proposed a project to facilitate the use of commercial cloud platforms for research teams to better understand cloud platform capabilities to support data-intensive research.
The project was designed to be competitive — six projects were awarded $100,000 each of cloud credits from AWS and GCP for a year to develop project workflows in the cloud. Internet2 managed the call for proposals and a peer-review process to select the six teams. Principal investigators were asked to nominate their preferred cloud platform.
After one year, teams were asked to present their projects at a workshop and provide a written report detailing their achievements using the cloud and how it helped advance their fields of science. They were then invited to submit proposals for the project’s second phase, where two teams would be awarded up to $500,000 each to be spent on cloud resources, and up to $330,000 each for staff and allowable in-direct costs. These proposals, together with phase one reports and presentations, were distributed for ad-hoc peer review.
We are now in phase two of the E-CAS project, which will be complete in October 2022. The two selected projects are “Deciphering the Brain’s Neural Code,” led by William Lytton and Salvador Dura-Burnal from SUNY Downstate Medical Centre, and “Heterogeneous computing for the Large Hadron Collider,” led by Philip Harris from MIT.
Through this project Internet2 and NSF are hoping to better understand:
- The extent to which commercial clouds can support large-scale computational science.
- The benefits of rapid or on-demand access to scalable resources and potentially large pools of new and specialist hardware, such as the latest generations of Graphic Processing Units (GPUs), Field Programmable Gate Arrays (FPGAs), and Application Specific Integrated Circuits (ASICs) such as Google’s Tensor-flow Processing Units (TPUs).
- The types of workflows that can benefit from Artificial Intelligence and Machine Learning in the cloud.
- Technical and support requirements for moving workflows from campus or publicly-funded facilities to commercial cloud platforms.
- The administrative issues in using commercial cloud including contract negotiations, account management, and financial processes and monitoring.
- Performance benefits or limitations of cloud compared to campus and national facilities.
To date, the cloud providers involved (AWS and GCP) have committed approximately $750,000 worth of cloud credits to the E-CAS project.,
For more detailed information on the E-CAS project, please visit our project space.
We want to hear from you.