Year: 2013
Type: conference-paper
Abstract: The first running period of the LHC was a great success. In particular vital for the timely analysis of the data were the good performance of the CMS detector and the well functioning distributed computing infrastructure. For the coming running period a significant increase of the data and event rates is expected, while for CMS there will be only a small increase in storage and CPU resources. In order to cope with this situation a very efficient usage of the resources is mandatory. CMS is moving to a more flexible model of data access, which explicitly uses access over the wide area network. This will allow a reduction of the necessary number of needed file replicas. In addition CMS is working on access to additional resources, such as Clouds and opportunistic sites that are available only for a limited duration. As a pilot project for this CMS is using its High Level Trigger farm during the LHC shutdown for processing via a Cloud interface. Recent developments in the CMS software stack are influenced by the trends of the hardware market. Therefore CMS is looking into multi-threaded applications as well as other hard architectures like ARM. Both have impact on the distributed computing infrastructure for job match making and scheduling. This presentation will report on the CMS experiences of the first LHC data taking period and will discuss the future developments.
Author: Wissing, Christoph
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7018029147