Automate migration to microservices architecture using Machine Learning techniques

Abstract: The microservice architectural style has many advantages such as scalability, reusability, and easy maintainability. Microservices have therefore become a popular architectural choice when developing new applications. Reaping these benefits requires redesigning the monolithic application and moving them to a microservices-based architecture. This process is inherently complex and costly, so automating this task is critical. This project proposes a method by which potential microservices can be identified from a given monolithic application while treating this problem as a clustering thematic. Our method takes as input the source code of the source application in order to apply different approaches to devise the one box application into its different microservices. In this report we detail each of these techniques while finishing with a discussion and comparison of the results.