Workflow Partitioning for Offline Distributed Execution on Mobile Devices

Abstract

Traditionally, workflow systems are built on the client/server architecture, in which a single workflow server takes the responsibility for the operation of the whole process, thereby requiring connections each time a task is completed. In cases where connection between client and server is not readily available - like in mobile environments, such an approach proves infeasible. Enabling the execution of a group of tasks by mobile clients in distributed and disconnected environments has been proposed as a possible solution. However, the partitioning of a workflow into groups of tasks for offline execution has not been adequately explored. This paper proposes an approach for workflow partitioning and an algorithm that enables automatic discovery of such partitions from a process model as a vital step in assigning grouped tasks. We have implemented the algorithm, evaluated and validated it and proposed ways in which it could be implemented in a real workflow environment.