Migration to microservices: a quality-driven approach

Microservices architecture has become increasingly popular among software practitioners in recent years as an effective approach to building complex applications that are more scalable, maintainable, and resilient. Top companies like Netflix, Amazon, and Uber have all successfully modernized their systems migrating them to microservices architecture. However, this migration process can pose certain challenges that require careful planning and execution to achieve desired outcomes and ensure that both functional and non-functional requirements are met. One of the most significant challenges in microservices migration is the planning phase, as poor planning can lead to increased complexity, reduced scalability, and degraded performance of the final system. To address these challenges, we propose a quality-driven migration approach that considers software qualities in all the migration stages. Our approach aims to improve software qualities with the migration process by applying architectural refactoring techniques, such as antipatterns detecion analysis. By combining techniques from both migration and refactoring, our approach can help organizations to achieve the desired outcomes of microservices migration while improving or maintaining quality attributes throughout the process. In addition, we created a novel quality-driven (antipatterns-based) refactoring approach to be applied to microservices derived from the migration. Given the industrial nature of the PhD of which this document represents the final Thesis, the research has been validated by applying it to a real case study from the BIM Italia company. In particular, the refactoring approach has been applied for refactoring two microservices of BIM Italia that suffered from significant performance degradation after migration. Industrial experimentation in BIM Italia has showcased the importance of quality-driven migration approaches in a microservices architecture. This Thesis makes several contributions, including: i) conducting a thorough analysis of quality-driven migration approaches, ii) proposing a quality-driven migration process for microservices that relies on antipatterns analysis to ensure a successful migration, iii) introducing a graph-based software representation with annotations for antipatterns detection, and iv) developing and validation of a quality-driven refactoring approach for microservices resulting from the migration of a monolithic system. The output of this Thesis is a set of guidelines for quality-driven migration to microservices that can help practitioners to avoid common pitfalls ensuring that their systems meet the expected quality requirements. Our work demonstrates the importance of considering quality attributes throughout the migration process and how architectural refactoring can help achieve these goals.