Operation-Oriented Model and Method for Effort Estimation of Architectural Migration Among Domain-Driven Design Variations

Abstract

Objective. The objective of this article is to improve the objectivity and accuracy of software effort estimation for architectural migration within the Domain-Driven Design (DDD) approach, specifically the transition from the Anemic Domain Model to the Rich Domain Model, by developing a comprehensive effort estimation method based on operational complexity. The proposed method is intended to overcome the subjectivity of traditional person-hour estimation and reduce the influence of the human factor.

Methodology. The method is grounded in an operational complexity model of business logic migration, which introduces two categories of migration operations – new logic introduction (N) and existing logic relocation (M) – along with a classification of base business logic migration operations, architectural classification of business logic by entity types, and distribution of migration operations across three layers: Application, Domain, and Infrastructure. The method extends the Fuzzy Use Case Size Point approach by introducing an extended Unadjusted Use-Case Size Points calculation model (xUUSP) with weighted parameters, and incorporates adaptive weight calibration using a modified gradient descent method.

Scientific novelty. The scientific novelty of this research lies in the development of an operation-oriented model and a method for estimating migration effort between Domain-Driven Design architectural variations. The proposed method is based on the extended Unadjusted Use-Case Size Points (xUUSP) model with weighted parameters and incorporates adaptive weight calibration using a modified gradient descent method.

Conclusions. As a result, the developed method provides higher objectivity and accuracy: on a dataset of 53 use cases across two Bounded Contexts, MMRE < 0.20 and PRED(0.25) ≥ 75–100% (depending on entity class) were achieved, exceeding standard prediction accuracy thresholds. In practice, the method can be applied to effort prediction, risk planning, and controlled evolutionary migration in enterprise systems with business logic built upon the Domain-Driven Design approach.