Checking soundness of business processes compositionally using symbolic observation graphs

Abstract

The Symbolic Observation Graph (SOG) associated with a labelled transition system and a subset of its labels is an efficient BDD-based abstraction representing the behavior of a system. The goal of this paper is to compose SOGs such that the resulting SOG is still small but represents the behavior of the composed business process in an appropriate way. In particular, we would like to deduce the properties of a composed business process by analysing the composition of the SOGs associated with its components. This question was already answered for the deadlock-freeness property in previous work. In this paper, we extend this result to other generic properties: the so-called soundness properties. These properties guarantee the absence of livelocks, deadlocks and other anomalies that can be formulated without domain knowledge. Thus, we show how the SOG can be adapted and used so that the verification of several variants of the soundness property can be performed modularly. © 2012 IFIP International Federation for Information Processing.