Towards Migrating Legacy Software Systems to Microservice-based Architectures: a Data-Centric Process for Microservice Identification

• Published in: 2022 IEEE 19th International Conference on Software Architecture Companion (ICSA-C) pp. 15 - 19
• Main Authors: Romani, Yamina, Tibermacine, Okba, Tibermacine, Chouki
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.03.2022
• Subjects: clustering; Computer architecture; Database-per-service pattern; Dictionaries; Microservice architectures; Microservices; monolithic to microservice migration; Scalability; Software architecture; Symbols; topic identification
• ISSN: 2768-4288
• DOI: 10.1109/ICSA-C54293.2022.00010

"Microservice-based architecture" is an architectural style exploited to develop software systems with the main concern of independent maintainability, deployability and scalability. These important capabilities in modern software development and operation settings led many companies to migrate their existing (legacy) monolithic software systems towards microservice-based architectures. The migration process is a challenging task. It requires splitting the system into consistent parts that represent the set of microservices. Existing works focus mainly on functional aspects in this splitting. We argue in this work that it would be beneficial to start this splitting by decomposing the database into clusters, where the data in each cluster is associated to a microservice's own independent database. This is commonly known as the "database-per-service" pattern in microservice architectures. This paper proposes our preliminary work on a data-centric process to identify microservices. This process performs database schema analysis and clustering in order to make topic identification. It aims at identifying a set of topics which correspond to potential microservices.