Switzerland is in a unique position to host one of the 4 off-site Data Centres (DC) due to its existing infrastructure for high-performance computing and data analysis at CSCS and experience in the Worldwide LHC Computing Grid (WLCG), from which CTAO inherits its data centre technologies. The Swiss off-site DC in Lugano will archive about half of the raw and simulated data. The data volume is expected to reach 50 Peta bytes by 2030. These big volumes will be processed at CSCS by the DPPS to produce high-level data products for use by astronomers, with the data production rate increasing as more telescopes join operations in the next few years.
The Swiss DC is already serving as the Off-site DC for the first operational Large Size Telescope of CTAO, for which a full archive of the raw data (close to 2 PetaBytes) has been transferred. CSCS is also providing the major part of the computing and storage for the production of the CTAO Monte-Carlo simulation and is setting up the computing environment for the first CTAO Science Data Challenge. In order to store and analyse this increasing data ow, both storage and compute backends were provided at CSCS exclusively for CTAO in a virtual Cluster code-named Vanil Noir, on top of the new Alps Infrastructure. These backends are deployed using modern software-defined infrastructure mechanisms and are built using Linux containers and Kubernetes.
At the beginning of 2024, the Alps Infrastructure will be expanded with a very significant amount of multi-GPUs nodes with NVIDIA Grace Hopper Superchips that will enable large, energy-efficient machine learning workflows such as neural network training and inference.

The Main components of the Swiss off-site DC: ARC and dCache middleware from WLCG receiving workload and data from RUCIO and DIRAC respectively, which in turn runs batch jobs and stores data at CSCS.