# Dynamic Virtualized Deployment of Particle Physics Environments on a High Performance Computing Cluster

## TL;DR

This paper presents a system for dynamically deploying virtualized particle physics environments on a high-performance computing cluster, enabling seamless, on-demand access for researchers working with LHC data.

## Contribution

It introduces a virtual machine-based deployment framework integrated with cloud orchestration and scheduling for particle physics research on HPC clusters.

## Key findings

- Effective virtualization of WLCG environments on NEMO cluster
- Automated resource management for virtualized and bare metal jobs
- Performance evaluation confirms suitability for particle physics applications

## Abstract

The NEMO High Performance Computing Cluster at the University of Freiburg has been made available to researchers of the ATLAS and CMS experiments. Users access the cluster from external machines connected to the World-wide LHC Computing Grid (WLCG). This paper describes how the full software environment of the WLCG is provided in a virtual machine image. The interplay between the schedulers for NEMO and for the external clusters is coordinated through the ROCED service. A cloud computing infrastructure is deployed at NEMO to orchestrate the simultaneous usage by bare metal and virtualized jobs. Through the setup, resources are provided to users in a transparent, automatized, and on-demand way. The performance of the virtualized environment has been evaluated for particle physics applications.