5th International ICST Conference on COMmunication System softWAre and middlewaRE

Research Article

Energy-efficient Workflow Distribution

  • @INPROCEEDINGS{10.1145/2016551.2016553,
        author={Daniel Fischer and Stefan F\o{}ll and Klaus Herrmann and Kurt Rothermel},
        title={Energy-efficient Workflow Distribution},
        proceedings={5th International ICST Conference on COMmunication System softWAre and middlewaRE},
        keywords={workflow distribution energy efficiency minimum cut},
  • Daniel Fischer
    Stefan Föll
    Klaus Herrmann
    Kurt Rothermel
    Year: 2012
    Energy-efficient Workflow Distribution
    DOI: 10.1145/2016551.2016553
Daniel Fischer1,*, Stefan Föll1, Klaus Herrmann1, Kurt Rothermel1
  • 1: Universität Stuttgart
*Contact email: daniel.fischer@ipvs.uni-stuttgart.de


Pervasive computing and business process modeling are increasingly joining forces, as mobile human users shall be seamlessly integrated into business processes. In respective scenarios, humans use mobile devices and wireless technology to interact with workflows running in a powerful back-end infrastructure. However, the frequent interaction between humans and workflows causes a high communication overhead and, thus, high energy consumption on mobile devices. This impacts the usability and efficiency of the business process due to rapidly drained batteries and the resulting short life-times of the devices and applications. We present an approach based on a minimum-cut algorithm for reducing costly data transmissions during workflow execution by distributing parts of a workflow to the users' devices. Our motivation is to reduce the energy consumption on the mobile devices and, thus, avoid draining batteries in the field. We prove that our algorithm finds the optimal solution for a given network and workflow, decreasing the energy consumed on mobile devices by 32-37% compared to an approach where the entire workflow is executed in the infrastructure. Thus, in typical domains like logistics and health care, one third of the energy can be saved. This either means that devices have to be charged less frequently, leading to less distraction in the business process, or that mobile device specifications can be lowered. Significant cost reductions result in both cases.