Automated enterprise architecture model maintenance via runtime IT discovery

Abstract

Enterprise architects struggle to cope with rapid architectural changes and to document them accordingly in their architecture models due to missing adequate tool support. For that reason, the documentation of Enterprise Architecture (EA) models is still mostly achieved by manual effort and often contains outdated information. In this work, we present a novel approach for EA model creation that leverages runtime service instrumentation of the existing IT architecture to automatically create, update, and enhance static EA models with runtime information. We introduce a new integration layer that synchronizes static and runtime data from different data sources. The hereby implemented prototype allows different stakeholders to explore information from both perspectives (static and runtime) based on a linked enterprise knowledge graph, which supports new use cases and analysis capabilities. We evaluate our prototype by implementing it in a big German retailer. The introduction of service naming conventions and validation workflows enables fully automated data integration, which minimizes the effort for manual tasks. Based on interviews we conducted with 17 experts from two different companies, we could prove that the tool is capable to automate EA model maintenance.