The evolving role of Tier2s in ATLAS with the new Computing and Data Distribution model

Abstract

Originally the ATLAS Computing and Data Distribution model assumed that the Tier-2s should keep on disk collectively at least one copy of all “active” AOD and DPD datasets. Evolution of ATLAS Computing and Data model requires changes in ATLAS Tier-2s policy for the data replication, dynamic data caching and remote data access. Tier-2 operations take place completely asynchronously with respect to data taking. Tier-2s do simulation and user analysis. Large-scale reprocessing jobs on real data are at first taking place mostly at Tier-1s but will progressively be shared with Tier-2s as well. The availability of disk space at Tier-2s is extremely important in the ATLAS Computing model as it allows more data to be readily accessible for analysis jobs to all users, independently of their geographical location. The Tier-2s disk space has been reserved for real, simulated, calibration and alignment, group, and user data. A buffer disk space is needed for input and output data for simulations jobs. Tier-2s are going to be used more efficiently. In this way Tier-1s and Tier-2s are becoming more equivalent for the network and the hierarchy of Tier-1, 2 is less strict. This paper presents the usage of Tier-2s resources in different Grid activities, caching of data at Tier-2s, and their role in the analysis in the new ATLAS Computing and Data model.