OpenShift-Based High Performance Computing for Research in Astrophysics

In this session, Project Vega project team members Nikolaos Moraitis, Software Engineer, Red Hat OpenShift team and Gabriel Szász,  PhD candidate, Masaryk University, Brno, discussed OpenShift-based High-Performance Computing for Research in Astrophysics. Studying the rotation of the stars led the research group at the Department of Theoretical Physics and Astrophysics (DTPA), Faculty of Science, Masaryk University, Brno, Czech Republic, to a technical problem.  The solution: use Kubernetes to provide the next-generation open-source tool for high-performance computing for their OpenShift cluster.

Abstract
Scientific computing as we know it today remains very conservative and researchers are often not keen to trust new technologies. This way of thinking often results in state-of-the-art research projects that provide excellent results, but lack scalability and proper software design. Project Vega uses existing codes for stellar rotation analysis, but these codes were not written with the cloud-native model in mind, and one simply cannot redesign them from scratch due to their complexity. 

This presented a challenge – determining how to conveniently run these pieces of the software inside a cloud-native environment. This idea originated from early communication with the Department of Theoretical Physics and Astrophysics at Masaryk University in Brno as part of a Ph.D. thesis on “New Pseudo-Rotating Stellar Atmospheres.” The thesis project has benefited from running in a modern distributed computing environment, and thanks to  Red Hat Research, we migrated the project to brand-new hardware in 2021. Different research groups are already interested in utilizing our OpenShift cluster. We are already working on covering a wider variety of workflows to satisfy their demands.

Speaker
Nikolaos Moraitis, Software Engineer in the Red Hat OpenShift Team

Conversation Leader
Gabriel Szász, PhD candidate at Masaryk University

Session Recording