Method for Migrating Legacy Monolithic Systems to Microservices Architecture

Abstract

Currently, most organizations depend on Information Systems. In general, these systems are built based on monolithic architecture, with centralized execution on a single server. Over the years, however, the constant changes to meet business needs and the accumulation of technical debt, have made these systems increasingly larger and complex, making aspects such as maintainability and scalability difficult. This difficult reality experienced by many organizations motivated the present work to investigate a new architectural trend called microservices. Popularized by companies like Netflix and Amazon, microservices can be an alternative for the modernization of legacy systems, providing low downtime and imperceptible impact to the end user. In this new architecture, the software is broken down into small parts that work independently and autonomously, bringing some improvements in terms of software quality attributes. However, there are also challenges and tradeoffs - adopting microservices tends to be a difficult process and not rarely unsuccessful, mainly due to the lack of methods to conduct the migration process. In this sense, based on scientific studies, this work presents a migration method entitled Microservice Full Cycle - MFC, inspired by the software development life cycle and DevOps strategies. The goal is to assist legacy software systems to gradually evolve guided by a set of steps and activities common to the microservice architecture. The validation of the MFC method is done through an experiment in a real application, showing advantages such as more cohesive and uncoupled codes, technological independence, agility in build, test and deploy (automation), scalability under high demand, greater interoperability and integration, geodistributed capabilities, plus realtime monitoring and feedback.