Flexible Persistence Support for State Machine-Based Workflow Engines

This paper introduces a flexible persistence model for state machine workflows. In contrast to the persistence support in contemporary process definition languages and corresponding workflow engines we propose a model where, firstly, an arbitrary number of persistence providers can be attached to the process engine and, secondly, the workflows can dynamically decide when they should be stored persistently and by which persistence provider. Furthermore, we exemplify the implementation of this model within ePVM, an embeddable process virtual machine based on the theoretical foundation of communicating extended finite state machines. Finally, we present results gained from an initial experiment demonstrating that the model can help increasing software efficiency.