Table of Contents
Fetching ...

Towards an Architecture-centric Methodology for Migrating to Microservices

Jonas Fritzsch, Justus Bogner, Markus Haug, Stefan Wagner, Alfred Zimmermann

TL;DR

The paper tackles the lack of a systematic path for migrating legacy monoliths to microservices by proposing an architecture-centric migration framework that structures the process into three phases and is supported by an accompanying tool. It synthesizes existing literature, taxonomy, and empirical insights from practitioner interviews and planned longitudinal case studies to enable architects to plan, reason about, and execute migrations with better decision-making and automation. The contributions include a holistic framework, a taxonomy-informed repository of techniques, and early tool support aimed at closing the academia-industry gap and improving knowledge transfer. The practical impact lies in providing actionable guidance for service identification, architectural refactoring, and iterative implementation, enabling more reliable and efficient modernization of large-scale systems.

Abstract

The euphoria around microservices has decreased over the years, but the trend of modernizing legacy systems to this novel architectural style is unbroken to date. A variety of approaches have been proposed in academia and industry, aiming to structure and automate the often long-lasting and cost-intensive migration journey. However, our research shows that there is still a need for more systematic guidance. While grey literature is dominant for knowledge exchange among practitioners, academia has contributed a significant body of knowledge as well, catching up on its initial neglect. A vast number of studies on the topic yielded novel techniques, often backed by industry evaluations. However, practitioners hardly leverage these resources. In this paper, we report on our efforts to design an architecture-centric methodology for migrating to microservices. As its main contribution, a framework provides guidance for architects during the three phases of a migration. We refer to methods, techniques, and approaches based on a variety of scientific studies that have not been made available in a similarly comprehensible manner before. Through an accompanying tool to be developed, architects will be in a position to systematically plan their migration, make better informed decisions, and use the most appropriate techniques and tools to transition their systems to microservices.

Towards an Architecture-centric Methodology for Migrating to Microservices

TL;DR

The paper tackles the lack of a systematic path for migrating legacy monoliths to microservices by proposing an architecture-centric migration framework that structures the process into three phases and is supported by an accompanying tool. It synthesizes existing literature, taxonomy, and empirical insights from practitioner interviews and planned longitudinal case studies to enable architects to plan, reason about, and execute migrations with better decision-making and automation. The contributions include a holistic framework, a taxonomy-informed repository of techniques, and early tool support aimed at closing the academia-industry gap and improving knowledge transfer. The practical impact lies in providing actionable guidance for service identification, architectural refactoring, and iterative implementation, enabling more reliable and efficient modernization of large-scale systems.

Abstract

The euphoria around microservices has decreased over the years, but the trend of modernizing legacy systems to this novel architectural style is unbroken to date. A variety of approaches have been proposed in academia and industry, aiming to structure and automate the often long-lasting and cost-intensive migration journey. However, our research shows that there is still a need for more systematic guidance. While grey literature is dominant for knowledge exchange among practitioners, academia has contributed a significant body of knowledge as well, catching up on its initial neglect. A vast number of studies on the topic yielded novel techniques, often backed by industry evaluations. However, practitioners hardly leverage these resources. In this paper, we report on our efforts to design an architecture-centric methodology for migrating to microservices. As its main contribution, a framework provides guidance for architects during the three phases of a migration. We refer to methods, techniques, and approaches based on a variety of scientific studies that have not been made available in a similarly comprehensible manner before. Through an accompanying tool to be developed, architects will be in a position to systematically plan their migration, make better informed decisions, and use the most appropriate techniques and tools to transition their systems to microservices.
Paper Structure (7 sections, 2 figures)

This paper contains 7 sections, 2 figures.

Figures (2)

  • Figure 1: Research Method
  • Figure 2: Proposed Framework for Microservices Migrations