CMS Experience with a World-Wide Data Federation

Not scheduled
15m
OIST

OIST

1919-1 Tancha, Onna-son, Kunigami-gun Okinawa, Japan 904-0495
poster presentation Track3: Data store and access

Speaker

Marian Zvada (University of Nebraska (US))

Description

Over the past three years, the CMS Collaboration has developed the “Any Data, Anytime, Anywhere” technology to make use of a global data federation that is based on the XrootD protocol. The federation is now deployed across virtually all Tier-1 and Tier-2 sites in the CMS distributed computing system. This data federation gives workflows greater flexibility for location of execution, which has benefits at all scales of operation, from individual users accessing specific collision event records to organized production processing. In preparation for the coming LHC data run, CMS has been testing a number of applications of the data federation at increasingly large scales. In this presentation, we will discuss the applications in use and the results of the tests.

Primary authors

Brian Paul Bockelman (University of Nebraska (US)) Dr Carl Vuosalo (University of Wisconsin (US)) Daniel Charles Bradley (University of Wisconsin (US)) Federica Fanzago (Universita e INFN, Padova (IT)) Frank Wuerthwein (Univ. of California San Diego (US)) Igor Sfiligoi (Univ. of California San Diego (US)) Kenneth Bloom (University of Nebraska (US)) Marian Zvada (University of Nebraska (US)) Matevz Tadel (Univ. of California San Diego (US)) Prof. Sridhara Dasu (University of Wisconsin (US)) Tommaso Boccali (Universita di Pisa & INFN (IT))

Presentation Materials