Managing variability across enterprise architecture layers: A structured literature review

Abstract

Digital transformation and the growing use of artificial intelligence increase the need for organisations to adapt business processes, data structures, applications, and technology platforms to different contexts. Enterprise architecture management (EAM) can help coordinate these elements, while variability modelling represents commonalities, alternatives, constraints, and dependencies between variants. However, research on variability has often concentrated on individual architectural layers, leaving cross-layer dependencies insufficiently addressed. This paper synthesises the structured literature review reported by Dehne and Sandkuhl and the supplied literature on EAM, feature modelling, and variability management. It compares representative approaches according to architectural layer, modelling technique, evaluation evidence, and principal limitation. The synthesis indicates that feature models, ArchiMate extensions, variability mining, model-driven engineering, and architecture description languages provide useful mechanisms, but no single approach in the reviewed set offers a validated, integrated method for managing variability across all enterprise architecture layers. The paper identifies cross-layer traceability, dependency representation, and evaluation as priorities for future work.