DIRAC—The Distributed MC Production and Analysis for LHCb

September 30, 2004 · One min read

A Tsaregorodtsev

Marseille, CPPM

P Charpentier

CERN

M Frank

CERN

V Garonne

Marseille, CPPM

M Witek

Cracow, INP

V Romanovski

Serpukhov, IHEP

U Egede

Imperial Coll., London

Vagnoni, V

INFN, Bologna

Korolko, I

Moscow, ITEP

Blouw, J

(Heidelberg, Max Planck Inst.) ;

Kuznetsov, G

(Rutherford)

Patrick, G

(Rutherford)

Gandelman, M

(U. Fed. Rio de Janeiro (main))

Graciani-Diaz, R

(U. Barcelona (main))

Bernet, R

(U. Zurich (main))

Brook, N

(Bristol U.)

Pickford, A

(Glasgow U.)

Tobin, M

(U. Liverpool (main))

Saroka, A

(U. Oxford (main))

Stokes-Rees, I

(U. Oxford (main))

Saborido-Silva, J

(U. Santiago de Compostela (main))

Sanchez-Garcia, M

(U. Santiago de Compostela (main)) Hide

DIRAC is the LHCb distributed computing grid infrastructure for Monte Carlo (MC) production and analysis. Its architecture is based on a set of distributed collaborating services. The service decomposition broadly follows the CERN/ARDA-RTAG proposal, which should allow for the interchange of the EGEE/gLite and DIRAC components. In this paper we give an overview of the DIRAC architecture, as well as the main design choices in its implementation. The light nature and modular design of the DIRAC components allows its functionality to be easily extended to include new computing and storage elements or to handle new types of tasks. The DIRAC system already uses different types of computing resources - from single PC's to a variety of batch systems and to the Grid environment. In particular, the DIRAC interface to the LCG2 grid will be presented.

More details: https://cds.cern.ch/record/1494103