Migrating Monolithic Web Applications to Microservice Architectures Leveraging Use Case and Component Similarity

Abstract

Microservice architectures have emerged as a solution to the challenges associated with monolithic systems, enabling better modularization, independent deployment, and improved maintainability. Identifying microservices in a legacy monolithic system is a critical step in the transformation of a monolithic system into a microservice architecture. Existing approaches to microservice identification often rely on dynamic analysis techniques that require system execution and complex dependency tracking, making them difficult to apply universally. This study proposes a novel semantics-based approach for microservice identification that leverages use cases and static source code analysis. By embedding both use cases and software component names using a sentence-transformer model, the proposed technique effectively classifies the components into microservices based on semantic similarity. Unlike existing methods, this approach does not require system execution, making it more efficient and applicable to various legacy systems. Experimental evaluations of open-source web applications demonstrate that the proposed method outperforms existing techniques in terms of accuracy and F1 score, thereby offering a reliable alternative for microservice identification.