Proceedings of the 10 th Advanced Summer School on Service Oriented Computing

In the field of process-aware information systems, much work has been devoted to developing metrics for determining the similarity between process models. Similarity assessment is important for a wide array of applications and one area that has received relatively little attention so far is the replaceability assessment of executable process models. Replaceability assessment is relevant during software migration, e.g., when upgrading to a new execution platform. In this setting, it might be necessary to replace process models that can no longer be executed on the newer platform. Many of the existing metrics are ill-suited for replaceability assessment, because they were developed for non-executable models and tend to abstract from details that are decisive for the aforementioned application scenario. This paper discusses existing metrics for similarity assessment in a literature review and selects a subset of the body of metrics as candidates for replaceability assessment. Using an exemplary computation, it recommends a particular metric, TAR-similarity, for this purpose. Through this evaluation, this paper can be seen as a motivation for developing better node mapping functions.