Detecting Data Inconsistency Based on Workflow Nets With Tables

Abstract

For a data element concurrently accessed by multiple threads or stored at different locations, improper data operations will cause its values to be inconsistent. This could damage the correctness of workflow systems. Limited by modeling method, most existing researches can only discuss this issue in the process data layer or the persistent data layer. And workflow nets with tables (WFT-nets) is a kind formal and conceptual model that is good at modeling data-flow from the both data layers. We opt for it to model workflow systems and detect inconsistent data. Firstly, based on the concepts of concurrency, data log and shared attribute, the error of data inconsistency is defined from the perspective of data value. Furthermore, a data refining strategy that is suitable for revealing data inconsistency is proposed. And then the reachable configuration graph (RC-graph) of a WFT-net can be constructed. Traversing it, all errors of data inconsistency can be checked. Finally, a set of comparative experiments shows the effectiveness of the proposed methods.