Computational Storage to Increase the Analysis Capability of Tier-2 HEP Data Sites

Chen Zou; Andrew A. Chien; Robert Gardner; Ilija Vukotic; IEEE Comp Soc

doi:10.1109/Cluster48925.2021.00066

Back

Conference proceeding

Computational Storage to Increase the Analysis Capability of Tier-2 HEP Data Sites

Chen Zou, Andrew A. Chien, Robert Gardner, Ilija Vukotic and IEEE Comp Soc

2021 IEEE International Conference on Cluster Computing (CLUSTER), Vol.2021-, pp.803-804

09/2021

DOI: https://doi.org/10.1109/Cluster48925.2021.00066

Abstract

Cluster computing

Computational storage

Conferences

Data centers

Energy storage

High energy physics

Large Hadron Collider

Large Hadron Collider (LHC) produces collision data at 100 PB/year which needs to be stored and analyzed for high energy physics (HEP) theories. We reconsider the design choices of HEP data centers and evaluate different upgrade options to improve their analysis capacity.Results show that computational storage to be the cost-effective and power-efficient upgrade option. Computational disks in the storage cluster deliver a 9.3-fold speedup for Higgs Boson analysis. This exceeds the speedup from all other upgrades considered (faster network: 100 to 1000 Gbps, upgrade from HDDs to SSDs).

Metrics

1 Record Views

Details

Title: Computational Storage to Increase the Analysis Capability of Tier-2 HEP Data Sites
Creators - without role: Chen Zou - University of Chicago
Andrew A. Chien - University of Chicago
Robert Gardner - Fermi National Accelerator Laboratory
Ilija Vukotic - Fermi National Accelerator Laboratory
IEEE Comp Soc
Publication Details: 2021 IEEE International Conference on Cluster Computing (CLUSTER), Vol.2021-, pp.803-804
Publisher: IEEE
Grant note: National Science Foundation (10.13039/100000001)
Identifiers: 9943502108331
Academic Unit: King Abdullah University of Science & Technology
Language: English
Resource Type: Conference proceeding