Distributed Data Management in Workflow Environments

Most existing workflow management systems (WFMSs) are based on a client/server architecture. This architecture simplifies the overall design but it does not match the distributed nature of workflow applicationsand imposes severe limitations in terms of scalability and reliability. Moreover, workflow engines are not very sophisticated in terms of data management, forgetting the fact that workflow is, to a great extent, data flow. In this paper, we propose a novel architecture to address the issue of data management in a WFMS. This architecture is based on a fully distributed workflow engine for control flow, plus a set of loosely synchronized replicated databases for data flow. The resulting system offers greater robustness and reliability as well as much better data handling capabilities than existing approaches. To better illustrate this novel architecture and its implications, two commercial systems are employed in this paper: FlowMark, as the workflow engine, and the replication capabilities of Lotus Notes, as the support system for distributed