Please use this identifier to cite or link to this item: http://bura.brunel.ac.uk/handle/2438/18786
Title: CMS Results of Grid-related activities using the early deployed LCG Implementations
Authors: Coviello, T
D'Amato, M
De Filippis, N
Donvito, G
Maggi, G
Pierro, A
Bonacorsi, D
Capiluppi, P
Fanfani, A
Grandi, C
Maroney, O
Nebrensky, JJ
Donno, F
Jank, W
Sciabà, A
Sinanis, N
Colling, D
Tallini, H
MacEvoy, BC
Wang, S
Kaiser, J
Osman, A
Charlot, C
Semenjouk, I
Biasotto, M
Fantinel, S
Corvo, M
Fanzago, F
Mazzucato, M
Verlato, M
Go, A
Chia Ming, K
Andreozzi, S
Cavalli, A
Ciaschini, V
Ghiselli, A
Italiano, A
Spataro, F
Vistoli, C
Tortone, G
Issue Date: 1-Sep-2004
Publisher: CERN
Citation: CMS Note, 2004
Abstract: The CMS Experiment is defining its Computing Model and is experimenting and testing the new distributed features offered by many Grid Projects. This report describes use by CMS of the early-deployed systems of LCG (LCG-0 and LCG-1). Most of the used features here discussed came from the EU implemented middleware, even if some of the tested capabilities were in common with the US developed middleware. This report describes the simulation of about 2 million of CMS detector events, which were generated as part of the official CMS Data Challenge 04 (Pre-Challenge-Production). The simulations were done on a CMS-dedicated testbed (CMS-LCG-0), where an ad-hoc modified version of the LCG-0 middleware was deployed and where the CMS Experiment had a complete control, and on the official early LCG delivered system (with the LCG-1 version). Modifications to the CMS simulation tools for events produc tion where studied and achieved, together with necessary adaptations of the middleware services. Bilateral feedback (between CMS and LCG middleware) played an important role in making progress (including bugs corrections). Fractions of successful processing of simulation jobs ranged from 70% to 90%. Most of the failure reasons were identified, with RLS instability being the greatest cause of failure. Evaluation of the LCG-1 middleware is also presented and discussed. The progress of the new functionalities introduced and the better-distributed organization of services were tested and eventually stressed. While the overall efficiency decreased in the early implementation of the system in respect of the LCG-0 testbed, a consistent success rate of 50-60% was achieved. One of the major difficulties for the simulation of CMS events on the LCG-1 system was identified to be the consistent configuration of the distributed sites.
URI: http://bura.brunel.ac.uk/handle/2438/18786
ISSN: CMS-NOTE-2004/034
CMS-NOTE-2004/034
Other Identifiers: CMS-NOTE-2004/034
CMS-NOTE-2004/034
Appears in Collections:Dept of Electronic and Electrical Engineering Research Papers

Files in This Item:
File Description SizeFormat 
FullText.PDF485.56 kBAdobe PDFView/Open


Items in BURA are protected by copyright, with all rights reserved, unless otherwise indicated.