KIT | KIT-Bibliothek | Impressum | Datenschutz

Workflow Generation with wfGenes

Roozmeh, Mehdi; Kondov, Ivan

Abstract (englisch):
Performing simulation and data analysis on su-percomputers requires using workflow management systems(WMSs). Due to the high diversity of application requirementsthere are currently many different WMSs in use with verydifferent input languages and enactment mechanisms. This hin-ders the reuse of workflows across these WMSs and makes thetransition from one WMS to another difficult. We propose alightweight framework based on a generative approach to makethis affordable for the WMS end user. With a proof-of-conceptimplementation, wfGenes, workflows for different workflowssystems are generated automatically from an abstract workflowdescription. We tested wfGenes for a use case in computationalnanoscience to generate workflows for FireWorks and SimStack. The framework can be readily extended to support further WMSsand to adopt a more standardized workflow description languagefor its input.

DOI: 10.1109/WORKS51914.2020.00007
Zugehörige Institution(en) am KIT Steinbuch Centre for Computing (SCC)
Publikationstyp Proceedingsbeitrag
Publikationsdatum 11.11.2020
Sprache Englisch
Identifikator ISBN: 978-0-7381-1040-0
KITopen-ID: 1000127830
HGF-Programm 46.11.01 (POF III, LK 01) Computational Science and Mathematical Methods
Erschienen in IEEE/ACM Workflows in Support of Large-Scale Science (WORKS)
Veranstaltung 15th IEEE/ACM Workflows in Support of Large-Scale Science (WORKS 2020), Online, 11.11.2020
Verlag Institute of Electrical and Electronics Engineers (IEEE)
Seiten 9-16
Projektinformation GRK 2450/1 (DFG, DFG KOORD, GRK 2450/1 - 2019)
Bemerkung zur Veröffentlichung Die Veranstaltung fand wegen der Corona-Pandemie als Online-Event statt
Schlagwörter scientific workflow, workflow management system, workflow description language, interoperability
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page