WLCG Mini-Capability Challenge: Host Tuning to Im- prove WAN Data Transfers

Abstract

High-throughput data movement is a defining requirement of the WLCG computing model, yet end-to-end performance is often constrained less by the network backbone than by host-level configuration. We present the results of a WLCG mini-capability challenge focused on host optimization at ATLAS and CMS sites (FNAL, UCSD, UNL, BNL, AGLT2, MWT2, and Vanderbilt), using EL8/9 hosts with 25\,Gbps or higher NICs. NET2 participated in a subsequent trans-Atlantic test. The methodology combines ESnet Fasterdata guidance with the OSG-hosted \texttt{fasterdata-tuning.sh} framework, applying TCP, queue, offload, and ring-buffer tuning with save and restore support for safe baseline comparisons. Results show that tuning benefit depends critically on the local bottleneck. Where DTNs and storage were both capable (UNL), FTS throughput rose from 60 to 90\,Gbps. Where storage was the binding constraint (UCSD), iperf3 throughput improved from 20 to 90\,Gbps while FTS was unchanged. Where the NIC was the binding constraint (Vanderbilt, 25\,G NIC), the tuning profile actively reduced throughput, motivating a review of per-site applicability. The challenge also surfaced and resolved a dCache proxying misconfiguration at AGLT2, demonstrating that mini-challenges serve a diagnostic function beyond raw performance measurement.

I Prove, Therefore I Am: Spatiotemporal Multi-Party Computation
Ziqing Guo, Fuyuki Kitagawa, Xiao Liang
arXiv22 Sep 2026PreprintCC BY-NC-ND 4.0Full textQuantum physics & quantum informationarXiv
Can the FCC-hh prove the B-L gauge symmetry?
Farinaldo S. Queiroz, J. Zamora-Saa, Ricardo C. Silva +1
The European Physical Journal C26 Oct 2026Peer-reviewedCC BY 4.0Full textHigh-energy & particle physicsCrossref
Many-Revolution Low-Thrust Transfers within Periodic Orbit Families
Ian M. Down
arXiv28 Aug 2026PreprintarXiv licenceFull textDynamical systems Statistical & nonlinear physicsarXiv
Octupole correlations in the excited 0⁺ states and two-neutron transfers in actinide nuclei
Kosuke Nomura
arXiv9 Oct 2026PreprintarXiv licenceFull textNuclear physicsarXiv
Inferring the Number of Quantum State Transfers with Neural Networks
Hayden Zammit, Tony J. G. Apollaro, Gianluca Valentino +1
arXiv28 Oct 2026PreprintCC BY 4.0Full textQuantum physics & quantum informationarXiv

Cite

BibTeX RIS

BibTeX

@misc{melo2026wlcg,
  title = {{WLCG Mini-Capability Challenge: Host Tuning to Im- prove WAN Data Transfers}},
  author = {Andrew Malone Melo and Carlos Fernando Gamboa and Diego Devila Foyo and Eduardo Bach and Eli Dart and Shawn McKee and Garhan Attebury and Philippe Laurens and Wendy Wu and Syed Asif Shaw and Hiro Ito},
  howpublished = {arXiv},
  year = {2026},
  month = sep,
  doi = {10.48550/arxiv.2609.35859},
  eprint = {2609.35859},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.35859},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - WLCG Mini-Capability Challenge: Host Tuning to Im- prove WAN Data Transfers
AU  - Andrew Malone Melo
AU  - Carlos Fernando Gamboa
AU  - Diego Devila Foyo
AU  - Eduardo Bach
AU  - Eli Dart
AU  - Shawn McKee
AU  - Garhan Attebury
AU  - Philippe Laurens
AU  - Wendy Wu
AU  - Syed Asif Shaw
AU  - Hiro Ito
PY  - 2026
DA  - 2026/09/25
PB  - arXiv
DO  - 10.48550/arxiv.2609.35859
UR  - https://arxiv.org/abs/2609.35859
AB  - High-throughput data movement is a defining requirement of the WLCG computing model, yet end-to-end performance is often constrained less by the network backbone than by host-level configuration. We present the results of a WLCG mini-capability challenge focused on host optimization at ATLAS and CMS sites (FNAL, UCSD, UNL, BNL, AGLT2, MWT2, and Vanderbilt), using EL8/9 hosts with 25\,Gbps or higher NICs. NET2 participated in a subsequent trans-Atlantic test. The methodology combines ESnet Fasterdata guidance with the OSG-hosted \texttt{fasterdata-tuning.sh} framework, applying TCP, queue, offload, and ring-buffer tuning with save and restore support for safe baseline comparisons. Results show that tuning benefit depends critically on the local bottleneck. Where DTNs and storage were both capable (UNL), FTS throughput rose from 60 to 90\,Gbps. Where storage was the binding constraint (UCSD), iperf3 throughput improved from 20 to 90\,Gbps while FTS was unchanged. Where the NIC was the binding constraint (Vanderbilt, 25\,G NIC), the tuning profile actively reduced throughput, motivating a review of per-site applicability. The challenge also surfaced and resolved a dCache proxying misconfiguration at AGLT2, demonstrating that mini-challenges serve a diagnostic function beyond raw performance measurement.
ER  -

Abstract

Related

Cite