Control-Flow-Based Methods to Support the Development of Sound Workflows

Workflows describe sequences of tasks to achieve goals. These sequences can contain decisions, loops, and parallelisations and are, therefore, similar to computer programs. Experts in the domain of workflow application usually design these workflows. However, these experts are rarely IT experts. For...

Full description

Bibliographic Details
Main Authors: Thomas M. Prinz, Wolfram Amme
Format: Article
Language:English
Published: Riga Technical University 2021-07-01
Series:Complex Systems Informatics and Modeling Quarterly
Subjects:
Online Access:https://csimq-journals.rtu.lv/article/view/4874
id doaj-81d5343738c34462b0b21fa287ad30d9
record_format Article
spelling doaj-81d5343738c34462b0b21fa287ad30d92021-08-10T07:41:25ZengRiga Technical UniversityComplex Systems Informatics and Modeling Quarterly2255-99222021-07-0102714410.7250/csimq.2021-27.012542Control-Flow-Based Methods to Support the Development of Sound WorkflowsThomas M. Prinz0Wolfram Amme1Course Evaluation Center, Friedrich Schiller University Jena, Am Steiger 3, Haus 1, 07743 JenaResearch group Program Analysis and Optimization, Friedrich Schiller University Jena, Ernst-Abbe-Platz 2, 07743 JenaWorkflows describe sequences of tasks to achieve goals. These sequences can contain decisions, loops, and parallelisations and are, therefore, similar to computer programs. Experts in the domain of workflow application usually design these workflows. However, these experts are rarely IT experts. For this reason, after automation by a computer, workflows can exhibit undesired behaviors. Such behaviors can be expensive and dangerous and should be avoided. The notion of soundness describes the absence of the undesired behaviors of deadlocks and abundances. The state of the art in workflow verification can detect such behaviors, but gives no indication of causes, does not provide detailed diagnostic information, or is slow. This article introducēs two new compiler-based techniques to find causes of deadlocks and abundances. These techniques provide detailed diagnostic information and have a cubic asymptotic complexity of runtime. Their efficiency and quality is evaluated using a benchmark of over thousand real-world workflows together with two leading state-of-the-art approaches.https://csimq-journals.rtu.lv/article/view/4874workflowverificationsoundnesscauses
collection DOAJ
language English
format Article
sources DOAJ
author Thomas M. Prinz
Wolfram Amme
spellingShingle Thomas M. Prinz
Wolfram Amme
Control-Flow-Based Methods to Support the Development of Sound Workflows
Complex Systems Informatics and Modeling Quarterly
workflow
verification
soundness
causes
author_facet Thomas M. Prinz
Wolfram Amme
author_sort Thomas M. Prinz
title Control-Flow-Based Methods to Support the Development of Sound Workflows
title_short Control-Flow-Based Methods to Support the Development of Sound Workflows
title_full Control-Flow-Based Methods to Support the Development of Sound Workflows
title_fullStr Control-Flow-Based Methods to Support the Development of Sound Workflows
title_full_unstemmed Control-Flow-Based Methods to Support the Development of Sound Workflows
title_sort control-flow-based methods to support the development of sound workflows
publisher Riga Technical University
series Complex Systems Informatics and Modeling Quarterly
issn 2255-9922
publishDate 2021-07-01
description Workflows describe sequences of tasks to achieve goals. These sequences can contain decisions, loops, and parallelisations and are, therefore, similar to computer programs. Experts in the domain of workflow application usually design these workflows. However, these experts are rarely IT experts. For this reason, after automation by a computer, workflows can exhibit undesired behaviors. Such behaviors can be expensive and dangerous and should be avoided. The notion of soundness describes the absence of the undesired behaviors of deadlocks and abundances. The state of the art in workflow verification can detect such behaviors, but gives no indication of causes, does not provide detailed diagnostic information, or is slow. This article introducēs two new compiler-based techniques to find causes of deadlocks and abundances. These techniques provide detailed diagnostic information and have a cubic asymptotic complexity of runtime. Their efficiency and quality is evaluated using a benchmark of over thousand real-world workflows together with two leading state-of-the-art approaches.
topic workflow
verification
soundness
causes
url https://csimq-journals.rtu.lv/article/view/4874
work_keys_str_mv AT thomasmprinz controlflowbasedmethodstosupportthedevelopmentofsoundworkflows
AT wolframamme controlflowbasedmethodstosupportthedevelopmentofsoundworkflows
_version_ 1721212474496122880