Work

CMS computing upgrade and evolution

Year: 2013

Type: conference-paper

Abstract: The distributed Grid computing infrastructure has been instrumental in the successful exploitation of the Large Hadron Collider data leading to the discovery of the Higgs boson. The computing system of the CMS experiment will need to face new challenges from 2015 on when LHC restarts with an anticipated higher detector output rate and event complexity, but with only a limited increase in the computing resources. A more efficient use of the available resources will be mandatory. CMS is improving the data storage, distribution and access as well as the processing efficiency. Remote access to the data through the Wide Area Network, dynamic data replication and deletion based on the data access patterns, and separation of disk and tape storage are some of the areas being actively developed. Multi-core processing and scheduling is being pursued in order to make a better use of the multi-core nodes available at the sites. In addition, CMS is exploring new computing techniques, such as Cloud Computing, to get access to opportunistic resources, or as a means of using without interference dedicated computing resources, such as the high level trigger farm, while they are not utilized for their main purpose. We discuss in this paper the ongoing work and plans to upgrade and evolve the CMS computing system.

Source: IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC)

DOI: 10.1109/nssmic.2013.6829578

Author: José M. Hernández

Institution: Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas

Language: English

Topic: Distributed and Parallel Computing Systems

Subfield: Computer Networks and Communications

Field: Computer Science

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: green

Funder: Science and Technology Facilities Council