Multi-View Spectral Clustering for Monolith-to-Microservices Migration
Gonzalo Rivera-Lazo, Fernando Montoya, Hernán Astudillo R. · 2025
The migration from monolithic architectures to microservices poses significant challenges, particularly in achieving modularity, decoupling and functional coherence. This paper explores a multi-view spectral clustering approach to systematically decompose monoliths by integrating structural dependencies, domain semantics and dynamic behavior into a unified weighted affinity model. The method uncovers optimal service boundaries that balance cohesion, coupling and complexity, generating multiple architectural options tailored to weighted quality attributes. Validation on a real-world application demonstrates how the framework enables flexible, quality-driven decompositions, supporting more informed and adaptive architectural decisions.