Author: Poeschl, M.C.
Paper Title Page
WEPOR051 Second Generation LHC Analysis Framework: Workload-based and User-oriented Solution 2784
  • S. Boychenko, C. Aguilera-Padilla, M.A. Galilée, J.C. Garnier, A.A. Gorzawski, K.H. Krol, J. Makai, M. Osinski, M.C. Poeschl, T.M. Ribeiro, A. Stanisz, M. Zerlauth
    CERN, Geneva, Switzerland
  • M.Z. Rela
    University of Coimbra, Coimbra, Portugal
  Consolidation and upgrades of accelerator equipment during the first long LHC shutdown period enabled particle collisions at energy levels almost twice higher compared to the first operational phase. Consequently, the software infrastructure providing vital information for machine operation and its optimisation needs to be updated to keep up with the challenges imposed by the increasing amount of collected data and the complexity of analysis. Current tools, designed more than a decade ago, have proven their reliability by significantly outperforming initially provisioned workloads, but are unable to scale efficiently to satisfy the growing needs of operators and hardware experts. In this paper we present our progress towards the development of a new workload-driven solution for LHC transient data analysis, based on identified user requirements. An initial setup and study of modern data storage and processing engines appropriate for the accelerator data analysis was conducted. First simulations of the proposed novel partitioning and replication approach, targeting a highly efficient service for heterogeneous analysis requests, were designed and performed.  
DOI • reference for this paper ※ DOI:10.18429/JACoW-IPAC2016-WEPOR051  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)