A Cloud+HPC hub for astrophysics and cosmology communities

The advancement of data-intensive analysis has emphasised the need for extensive, scalable environments that enable easy access to diverse computational resources. This project proposes the development of a Scientific Hub for the astrophysics and cosmology communities, leveraging a global, cloud-based infrastructure that integrates Kubernetes-based offloading to ensure efficient, user-friendly access to High Performance Computing (HPC), High Throughput Computing (HTC) and cloud resources. This solution introduces “transparent offloading” of containerized workloads using Kubernetes API primitives, creating a unified cloud-native interface that abstracts complexities and optimally allocates resources based on the specific needs of each task. Through the Scientific Hub, users could access JupyterHub-supported environments with configurable resources, persistent data storage and secure authentication, facilitating the execution and sharing of scientific workflows in a robust, scalable manner. This contribution showcases the aims and challenges of the development of the Scientific Hub on ICSC partners’ resources (including the Leonardo supercomputer at CINECA), along with some key use cases.