A pattern-based approach to quantitative enterprise architecture analysis

Enterprise Architecture (EA) management involves tasks that substantially contribute to the operations of an enterprise, and to its sustainable market presence. One important aspect of this is the availability of services to customers. However, the increasing interconnectedness of systems with other systems and with business processes makes it difficult to get a clear view on change impacts and dependency structures. While management level decision makers need this information to make sound decisions, EA models often do not include quality attributes (such as availability), and very rarely provide quantitative means to assess them. We address these shortcomings by augmenting an information model for EA modeling with concepts from Probabilistic Relational Models, thus enabling quantitative analysis. A sample business case is evaluated as an example of the technique, showing how decision makers can benefit from information on availability impacts on enterprise business services.