Speakers
Description
INFN manages the DATAcloud infrastructure, a federated and scalable network of cloud computing sites. Within the framework of the ICSC project (National Research Center in High-Performance Computing, Big Data, and Quantum Computing), funded by the Italian National Recovery and Resilience Plan (PNRR), research and development activities are carried out to foster innovation in high-performance computing, large-scale simulation, and big data analytics. In this context, the INFN Milano computing center has expanded its computing capacity through the deployment of a bare-metal Kubernetes cluster.
This contribution describes the deployment of an HTCondor cluster on Kubernetes to support Container Universe workloads using both Docker and Apptainer runtimes. HTCondor execute nodes are run as containers and managed as Kubernetes Deployments, enabling dynamic scaling and integration with an existing bare-metal HTCondor cluster. Monte Carlo production workflows have been executed in multiple configurations, including Apptainer jobs on bare metal, Apptainer and Docker-based jobs running on Kubernetes, and Docker-based HTCondor deployments on virtual machines managed by OpenStack. The results show that the adoption of virtualization and container orchestration layers does not introduce significant performance degradation for the evaluated workloads, while providing increased flexibility and portability of the computing infrastructure.