Speaker
Description
CloudVeneto is a private cloud targeted to scientific communities, based on OpenStack software, designed in 2013 to support INFN projects, initially mostly Nuclear Physics and HEP ones. During the last 12 years it evolved by integrating resources and use cases of several Departments of the University of Padova. It currently supports several scientific disciplines of different domains, but it is also used in other scenarios (e.g. to support teaching activities).
In 2025 CloudVeneto almost doubled its computing power capacity by integrating an “HPC Bubble”, i.e. a high-performance cluster designed for intensive computations, combining CPU and GPU nodes with low-latency interconnections, funded by the Italian National Recovery and Resilience Plan (NRRP). The Bubble is managed through a SLURM-based batch system, which enables flexible workload management. Given the features offered and the ease of use, several groups have decided to fund additional resources for the cluster. The Bubble is currently used by users involved in NRRP projects and in an opportunistic mode also by other users. Despite the HPC cluster being operated by a batch system outside the cloud environment, it shares with CloudVeneto a set of services as: resource provisioning and configuration; Ceph-based storage cluster; resource monitoring. Moreover, upon users’ request we can easily reconfigure some HPC nodes into Virtual Machines in the CloudVeneto environment, and viceversa.
In this paper we’ll describe in detail the solutions adopted to quickly put in production the HPC Bubble resources in an effective and integrated manner with the existing CloudVeneto infrastructure.