Multilevel complexity measurement in enterprise architecture models

Abstract

The authors create a model to measure the complexity associated with enterprise architecture (EA) proposals at the design stage. At this stage, existing methods measure the complexity of business capabilities regardless of their leveraging business and technical architectures. Measuring their complexity at the implementation level restricts product changes and generates higher costs. Our proposed model analyses the detailed EA design, thereby improving enterprise modelling and integration. First, we define the EA functionalities and dependencies to measure the complexity of a project roadmap. Next, we extend an existing mathematical formula to measure complexity in terms of these functionalities and dependencies at three additional and increasingly granular levels of detail: gap analysis model, target architecture design and implementation architecture design. We apply our multilevel complexity measurement to EA case studies. The results show that our model provides a more accurate complexity measurement compared with existing methods; the complexity effort estimated at the implementation design level has an accuracy between 63% and 94% with respect to the actual roadmap’s implementation effort. Our method provides an early predictor of EA reliability, which could enable organisations to better control project scope, time and budget, especially in reuse-based implementations such as service-oriented architectures.