News

At the recent International Supercomputing Conference (ISC) in Hamburg, Germany, Christine Kirkpatrick, Research Data Services director at the San Diego Supercomputer Center (SDSC) at UC San Diego, participated in two Birds of a Feather (BoF) discussions (gatherings among like-minded colleagues to discuss topics of shared interest): Bridging the HPC/Data Divide: Data Challenges and Solutions in the HPC World and A National Science Data Fabric to Democratize Data Access and Reusability.

In Bridging the HPC/Data Divide: Data Challenges and Solutions in the HPC World, Kirkpatrick and a team of international experts continued conversations from the data and high performance computing (HPC) communities held at the Supercomputing Conference 22 in Dallas, Texas, and International Data Week in Seoul, Korea. In this BoF, they addressed specific data challenges and solutions in the HPC arena and discussed paths for closer collaboration in the future. The BoF speakers included Kirkpatrick, Mark Gray and Maciej Cytowski from the Pawsey Supercomputing Centre in Australia, Marek Michalewicz from the Sano Centre for Computational Personalised Medicine - International Research Foundation, as well as Maijastiina Arvola and Hanna Koivula from the CSC - IT Center for Science in Finland. The discussion specifically probed topics such as data management techniques, including robust vocabularies and new methods of producing, curating and mining HPC-generated data.

In the session A National Science Data Fabric to Democratize Data Access and Reusability, Kirkpatrick engaged the data science community in discussion about the challenges and opportunities of the National Science Data Fabric (NSDF) project and other similar efforts to connect an open network of institutions, including resource-disadvantaged institutions. She worked with Jay Lofstead from the Sandia National Laboratories in New Mexico, Valerio Pascucci from the Pacific Northwest National Laboratory in Washington, and Michela Taufer and Jakob Luettgau from the Global Computing Laboratory at University of Tennessee Knoxville to report on NSDF achievements aiming to develop a data fabric that can take advantage of economies to fill the missing middle in our current computational infrastructure. In this BoF, the group discussed an array of opportunities for collaboration between the NSDF project and international organizations.

“Infrastructure by its nature is very integrative,” Kirkpatrick said. “Decades ago one might focus on only one aspect of computation, but to address big science questions we must work across HPC, cloud and storage platforms—a variety of open source frameworks—all with a deep understanding of the underlying use cases and data challenges specific to the domain. This will require a more tightly coupled data fabric like NSDF, and closer collaboration from many types of experts, including HPC and data people.”