Publication on the smart job scheduler implementation

Abstract:

Project: EuroScienceGateway was funded by the European Union programme Horizon Europe (HORIZON-INFRA-2021-EOSC-01-04) under grant agreement number 101057388 and by UK Research and Innovation (UKRI) under the UK government’s Horizon Europe funding guarantee grant number 10038963.

Document: D4.2 Publication on the smart job scheduler implementation

Work Package: Work Package 4. Building blocks for a sustainable operating model.

Task:

  • Task 4.3 Implement a smart job-scheduling system across Europe

Lead Beneficiary: EGI

Contributing Beneficiary: ALU-FR, CESNET, EGI, UiO, and VIB

Executive Summary

Galaxy is currently using the Total Perspective Vortex (TPV) to schedule millions of jobs for hundred thousand users globally. While TPV has proven to be a robust meta-scheduling tool for Galaxy in the last years, there are areas of improvement that have been addressed in the EuroScienceGateway project:

  • Gathering live usage metrics from across the distributed computing endpoints connected to Galaxy in order to distribute the load across all sites.
  • Adding latitude and longitude attributes to data stores and computing endpoints to allocate jobs as close as possible to the location of the data.
  • Visualizing job distribution across sites with an intuitive dashboard.

As a result the EuroScienceGateway project has developed two new tools:

  • TPV Broker for the efficient meta-scheduling of jobs taking into account real-time usage metrics and data-locality information
  • Galaxy Job Radar: a web dashboard to easily visualize the allocation of jobs across all sites

The EuroScienceGateway project has significantly improved the meta-scheduling of jobs for Galaxy, resulting in less waiting times for users to see their job completed and improving resource utilization across all sites.

SEEK ID: https://workflowhub.eu/publications/52

DOI: 10.5281/zenodo.15827107

Teams: EuroScienceGateway

Publication type: Tech report

Publisher: Zenodo

Citation: https://doi.org/10.5281/zenodo.15827107

Date Published: 27th Feb 2025

URL: https://doi.org/10.5281/zenodo.15827107

Registered Mode: imported from a bibtex file

Authors: Abdulrahman Azab, Sanjay Kumar Srikakulam, Paul De Geest, Tomáš Vondrák, Björn Grüning, Mira Kuntz, Enol Fernandez-del-Castillo, Sebastian Luna-Valero

Citation
Azab, A., Srikakulam, S. K., De Geest, P., Vondrák, T., Grüning, B., Kuntz, M., Fernandez-del-Castillo, E., & Luna-Valero, S. (2025). Publication on the smart job scheduler implementation. Zenodo. https://doi.org/10.5281/ZENODO.15827107
Activity

Views: 55

Created: 3rd Oct 2025 at 17:05

help Tags

This item has not yet been tagged.

help Attributions

None

Powered by
(v.1.17.0-main)
Copyright © 2008 - 2025 The University of Manchester and HITS gGmbH