A Comparative Study of Meta-Data-Based Microservice Extraction Tools

Abstract

Microservices are state-of-the-art architectural patterns for developing highly resilient and scalable applications to be hosted on the web. One of the tedious and challenging aspects on applying microservice pattern is the process of transforming legacy "monolithic" software to microservices. This process, named microservice extraction, is mostly done manually because of its complicated nature and the need for domain expertise. In this study, meta-data driven tools that aim to perform automatic extraction of microservices are focused. Initially, the studies that perform automatic microservice extraction are reviewed to identify public meta-data based tools. Afterwards, the identified tools are investigated by their relative success in extracting microservices. Reviewed tools suffer over-decomposition during extraction and most of them fail to provide a mature basis when considered to exhibit practical value to be used in real-world cases.