Semantic Normalization and Merging of Business Dependency Models

Assessing potential threats and impacts relevant for a company, requires a detailed analysis of a company's business processes and functions down to a level of infrastructure resources, in the form of one business dependency model. Required information is frequently encapsulated in BPMN models per process, but pose an eminent problem of fusing and merging multiple sources into one model. Experts defining BPMN models possibly use different nomenclature, descriptions, and references towards common entities, leading to semantically overlapping partial dependency models. Merging multiple partial dependency models is a novel problem related to the business process matching problem, but origins from an orthogonal perspective. In this paper we propose a business dependency model normalization and matching approach by exploiting structures and dependencies of business resources, which neither requires linguistic processing nor "fuzzy" matching processes.