Compute and Storage Infrastructural Challenges Towards Promotion and Acceleration of Data-driven Research at the University of Osaka

Susumu Date (The University of Osaka)

Abstract

AI for Science highly requests a flexible set of compute and data resources to allow researchers to perfort data-driven scientific research. For the reason, D3 Center at The University of Osaka has been working on the development of a Cyberinfrastructure that allow researchers to efficiently apply HPC and AI techniques onto scientific data sets. As of writing this, to this end, we have operated ONION as data aggregation infrastructure on top of compute resources such as SQUID and OCTOPUS as supercomputing systems and mdxII as a cloud infrastructure. For further advancement, we have been working on the realization of RED-ONION as high-speed data transfer service. In this talk, the speaker will talk about the overview of compute and storage infrastructure which we envisage as well as challenges we face.

<= Go back