Cost analysis for a resource sensitive workflow modelling language

Abstract

Workflow analysis usually requires domain-specific knowledge from the domain experts, making it a relatively manual process. In addition, workflows often cross organisational boundaries. As a result, minor local modifications in the workflow of a collaborative partner may be propagated to other concurrently running tasks of the workflow, which is difficult for the domain experts to recognise since they only have a limited (local) view of the workflow. Therefore, changes in cross-organisational workflows may result in significant adverse impacts. This paper presents a resource-sensitive formal modelling language, , which has explicit notions of task dependencies, qualitative assessment of resources, time advancement and method execution deadlines. The language allows the workflow analysers to estimate the effect of changes in collaborative workflows with respect to cost in terms of execution time. This paper proposes a static analysis to compute the worst execution time of a cross-organisational workflow modelled in by defining a compositional function that translates an program to a set of cost equations. • In most cases, workflow analysis requires domain-specific knowledge, which makes it a relatively manual process. Additionally, workflows often cross organizational boundaries. Thus, slight local modifications in the workflow of a collaborative partner may be propagated to other concurrently running tasks of the workflow, which is difficult for domain experts to discern since they have only a limited (local) view of the workflow. Therefore, changes in cross-organizational workflows could have adverse effects. This paper presents a resource-sensitive formal modelling language, Rpl. The language has explicit notions for task dependencies, resource allocation and time advancement. The language allows the workflow analyzers to estimate the effect of changes in collaborative workflows with respect to cost in terms of execution time. This paper proposes a static analysis for computing the worst execution time of a cross-organizational workflow modelled in Rpl by defining a compositional function that translates an Rpl program to a set of cost equations.