Skip to main navigation Skip to search Skip to main content

Serverless Workflow Benchmarking with HSMFlow and WfBench

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The convergence of edge computing, big data analytics, and artificial intelligence (AI) with traditional scientific calculations is increasingly being adopted in high-performance computing (HPC) workflows. Workflow management systems are crucial for managing and orchestrating these complex computational tasks. However, it is difficult to identify patterns within the growing population of HPC workflows. Moreover, serverless has emerged as a novel computing paradigm, offering dynamic resource allocation, quick response time, fine-grained resource management and auto-scaling. In this tutorial, we present a framework to evaluate serverless executions of HPC scientific workflows. Our approach integrates a widely used traditional HPC workflow generator with an HPC serverless workflow management system to create benchmark suites of scientific workflows with diverse characteristics.

Original languageEnglish
Title of host publicationMIDDLEWARE Tutorials 2025 - Proceedings of the 2025 International Middleware Conference Tutorial Track, Part of Middleware 2025
PublisherAssociation for Computing Machinery, Inc
Pages3-6
Number of pages4
ISBN (Electronic)9798400723209
DOIs
StatePublished - Dec 14 2025
Event2025 International Middleware Conference Tutorial Track, Middleware Tutorials 2025 - Nashville, United States
Duration: Dec 15 2025Dec 19 2025

Publication series

NameMIDDLEWARE Tutorials 2025 - Proceedings of the 2025 International Middleware Conference Tutorial Track, Part of Middleware 2025

Conference

Conference2025 International Middleware Conference Tutorial Track, Middleware Tutorials 2025
Country/TerritoryUnited States
CityNashville
Period12/15/2512/19/25

Fingerprint

Dive into the research topics of 'Serverless Workflow Benchmarking with HSMFlow and WfBench'. Together they form a unique fingerprint.

Cite this