Optimizing large data transfers for the ALICE experiment in Run 3

The physics programme and scope of HEP experiments natu-rally grow with time, thus the computing requirements, both CPU and storage, increase. The 4 large LHC experiments (ALICE, ATLAS, CMS, and LHCb) are undergoing upgrade cycles, for ALICE and LHCb the upgrade was finalized in the years of the LHC long Shutdown 2, 2019–2021, and for ATLAS and CMS it is foreseen to take place in 2026–2028. The ALICE upgrade was characterized by a complete overhaul of the detector, data acquisition (DAQ) systems, and entire software stack. As a consequence, the experiment can take up to 100 times more events, compared to the previous setup, and is employing a new online data compression and calibration utility (O2) to reduce the data stream, which nonetheless is 4 times larger than before the upgrade. To manage the new more complex data paths, a new data management software was designed and deployed as part of the Grid software upgrade. This article present the ALICE data movement system from the data compression farm O2 to the different storage instances around the world, with a focus on transfer optimization and automating data transfers.