Speaker
Dr
Robert W Gardner
(University of Chicago)
Description
DRAFT
We discuss the US ATLAS experience in deploying, integrating and operating a set of distributed computing centers into a single, logical computing facility suitable to meet the simulation and analysis requirements for the first years of LHC physics. Particular attention is paid to the Tier 2 layer in the LHC computing hierarchy, which is a focal point of activity supporting centrally managed tasks from CERN, managed data transfers to and from the Tier 1 center at Brookhaven National Laboratory as well as Tier 1 and Tier 2 centers in Europe and Asia, sporadic user-driven transfers to university-based Tier 3 clusters, as well as chaotic analysis activity from a globally distributed collaboration. The Tier 2 centers provide a resource at an intermediate scale in HEP: larger than typical physics departmental computing clusters but smaller than facilities at the national laboratories, and therefore present a unique set of challenges. We describe new challenges emerging as the computing model evolves to meet the needs of ATLAS physics, implying Tier 2 computing scales reaching multiple-Petabyte storage systems and many-thousand computing cores.
Primary author
Dr
Robert W Gardner
(University of Chicago)
