AI-based Quality-driven Decomposition Tool for Monolith to Microservice Migration
Muhammad Hafiz Hasan, Mohd Hafeez Osman, Novia Indriaty Admodisastro, Sufri Muhammad · 2023
Businesses and organizations are increasingly moving their business-critical systems to the cloud environment to take advantage of cloud benefits, thus improving its agility, maintainability, and flexibility. Businesses decompose their legacy monolith applications to cloud-native architecture such as microservice to leverage the cloud-native potential. With the expanding distributed nature of migrated applications, new challenges in determining the migrated architecture quality arise. Previous migration framework gives minimal attention to the post-migration quality aspect. This paper presents the Structural Quality (S-Quality) tool - a quality-driven decomposition tool that uses machine learning for migrating monolith applications to the microservice architecture. This tool utilizes various clustering techniques on monolith structural design properties to determine the service boundaries. In addition, this tool facilitates identifying the best microservice candidates based on migration architectural quality objectives through the scoring algorithm method, hence another contribution of this work. We validate our developed tool and scoring algorithm using a semi-structured interview with experts from the industry. Overall, findings indicate that the proposed scoring algorithm shows positive feedback from the experts and the acceptance of the S-Quality tool applicability by the industries.