CMS The Computing Project

Acknowledgements The CMS Computing Group gratefully acknowledges the contributions of technical staff throughout the world who have been involved in the design, operations and analysis of the computing challenges that have led us to this TDR. The CMS computing technical design has been developed in cooperation with our colleagues in the Worldwide LHC Computing Grid, together with the computing teams of ALICE, ATLAS and LHCb. We thank them all for their collaboration and for their assistance with and operation of the data-challenges that form the underpinning of this report. We would like to thank Neil Geddes who has diligently followed the drafts of this TDR and contributed many important suggestions. We thank also the external reviewers of our Computing Model Paper: Neil, Tony Cass and John Harvey. We thank our CMS internal reviewers Jim Branson, Bob Clare, Lorenzo Foa and Gail Hanson for all their help and constructive comments. For their perpetual good humour in the face of our unreasonable requests and deadlines, we thank the CMS Secretariat: Special thanks to Sergio Cittolin for his artistic interpretation of the CMS Computing Model shown on the cover page We also wish to thank our collaborators on CMS and especially the CMS management for their continuous support and encouragement. xix xx Executive Summary This document provides a top-level description of the organisation of the CMS Offline Computing systems. This organisation relies upon a number of cooperating pieces: • A tier-organised structure of computing resources, based on a Tier-0 centre at CERN and a small number of Tier-1 centres connected using high-speed networks. • A relatively large number of Tier-2 analysis centres where physics analysis will be performed. • A comprehensive and performant software framework designed for high-energy event streams. • Workload management tools to coordinate work at the centres and data management tools to ensure the efficient use of computing resources and the integrity of the data, including adequate auditing and safekeeping of raw and reconstructed data, calibration data, and job parameters. • A comprehensive project management plan so that the various project deliver-ables are tracked and potential bottlenecks are detected and eliminated. These pieces are discussed in this document in terms of the CMS Data Model and the Physics Analysis Models. The workload management is considered in the context of integration into the LHC Computing Grid. Chapter 2 provides the motivation behind the top-level baseline CMS Computing Model and describes it …