Speaker
Description
The ICSC initiative (Italian National Centre for High-Performance Computing, Big Data, and Quantum Computing) is creating a flexible cloud platform to manage the escalating computational requirements of the High-Luminosity Large Hadron Collider (HL-LHC) and future collider projects. This approach leverages Kubernetes for orchestration and containerized deployments to streamline access to heterogeneous computing resources. It empowers users to execute interactive and near-real-time analytical tasks via platforms like JupyterHub and Dask, while abstracting away the complexities of the supporting infrastructure.
To evaluate the platform's functionality and efficiency, a comprehensive benchmarking initiative was conducted employing authentic workloads derived from high-energy physics studies. These evaluations examined factors such as throughput, adaptability, and user-friendliness across varied setups. Afterward, a parallel bare-metal system was configured with the same software suite and testing protocols, facilitating direct performance contrasts between cloud and on-site environments.
The research delivers early perspectives on the balances between efficiency and scalability in various computing frameworks, while advancing the creation of consistent, replicable benchmarks for advanced HPC and data-driven processes in particle physics.