Design and performance of the virtualization platform for offline computing on the ATLAS TDAQ Farm

20th International Conference on Computing in High Energy and Nuclear Physics

物理学;计算机科学

Ballestrero, S.^1 ; Batraneanu, S.M.^2 ; Brasolin, F.^3 ; Contescu, C.^4,5 ; Girolamo, A. Di^4 ; Lee, C.J.^1,4 ; Astigarraga, M. E. Pozo^4 ; Scannicchio, D.A.^2 ; Twomey, M.S.^6 ; Zaytsev, A.^7

University of Johannesburg, South Africa^1

University of California, Irvine, United States^2

Istituto Nazionale di Fisica Nucleare Sezione di Bologna, Italy^3

CERN, Switzerland^4

Polytechnic University of Bucharest, Romania^5

University of Washington, Department of Physics, United States^6

Brookhaven National Laboratory (BNL), United States^7

关键词: ATLAS experiment;    Cloud managements;    Computing resource;    MC simulation;    Off-line computing;    Private networks;    Processing activity;    Production jobs;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/513/3/032011/pdf
DOI  :  10.1088/1742-6596/513/3/032011

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

With the LHC collider at CERN currently going through the period of Long Shutdown 1 there is an opportunity to use the computing resources of the experiments' large trigger farms for other data processing activities. In the case of the ATLAS experiment, the TDAQ farm, consisting of more than 1500 compute nodes, is suitable for running Monte Carlo (MC) production jobs that are mostly CPU and not I/O bound. This contribution gives a thorough review of the design and deployment of a virtualized platform running on this computing resource and of its use to run large groups of CernVM based virtual machines operating as a single CERN-P1 WLCG site. This platform has been designed to guarantee the security and the usability of the ATLAS private network, and to minimize interference with TDAQ's usage of the farm. Openstack has been chosen to provide a cloud management layer. The experience gained in the last 3.5 months shows that the use of the TDAQ farm for the MC simulation contributes to the ATLAS data processing at the level of a large Tier-1 WLCG site, despite the opportunistic nature of the underlying computing resources being used.