Total Cost of Ownership and Evaluation of Google Cloud Resources for the ATLAS Experiment at the LHC

Autor
Aad, G.
Aakvaag, E.
Abbott, B.
Abdelhameed, S.
Abeling, K.
Abicht, N. J.
Abidi, S. H.
Aboulhorma, A.
Abramowicz, H.
Abreu, H.
Baroň, Petr
Biroš, Marek
Chauhan, Chainika
Divíšek, Martin
Federič, Pavol
Haviernik, Matej
Hidić, Nihad
Huang, Yicong
Jankových, Samuel
Karkošová Martinovicová, Gabriela
Pauwels, Franciszek Jerzy Monique
Petrů, Tadeáš
Pleskot, Vojtěch
Poláček, Stanislav
Rybář, Martin
Sýkora, Martin
Todorova, Šárka
Váňa, Pavel
Datum vydání
2025Publikováno v
Computing and Software for Big ScienceNakladatel / Místo vydání
Springer International PublishingRočník / Číslo vydání
2025 (9)ISBN / ISSN
ISSN: 2510-2036ISBN / ISSN
eISSN: 2510-2044Informace o financování
UK//COOP
MSM//SVV260829
MSM//EH22_008/0004632
MSM//UNCE24/SCI/016
MSM//LM2023040
Metadata
Zobrazit celý záznamKolekce
Tato publikace má vydavatelskou verzi s DOI 10.1007/s41781-024-00128-x
Abstrakt
The ATLAS Google Project was established as part of an ongoing evaluation of the use of commercial cloudsby the ATLAS Collaboration, in anticipation of the potentialfuture adoption of such resources by WLCG grid sitesto fulfil or complement their computing pledges. Seamlessintegration of Google cloud resources into the worldwideATLAS distributed computing infrastructure was achievedat large scale and for an extended period of time, and hencecloud resources are shown to be an effective mechanism toprovide additional, flexible computing capacity to ATLAS.For the first time a total cost of ownership analysis has beenperformed, to identify the dominant cost drivers and exploreeffective mechanisms for cost control. Network usage significantlyimpacts the costs of certain ATLAS workflows,underscoring the importance of implementing such mechanisms.Resource bursting has been successfully demonstrated,whilst exposing the true cost of this type of activity.A follow-up to the project is underway to investigate methodsfor improving the integration of cloud resources in dataintensivedistributed computing environments and reducingcosts related to network connectivity, which represents the primary expense when extensively utilising cloud resources.
Klíčová slova
Cost, Cloud resources, ATLAS detector, LHC, CERN
Trvalý odkaz
https://hdl.handle.net/20.500.14178/3550Licence
Licence pro užití plného textu výsledku: Creative Commons Uveďte původ 4.0 International
