Multivocal study on microservice dependencies

doi:10.1016/j.jss.2025.112334

2025-04-01

PRE

Amr S. Abdelfattah

Tomáš Černý *

Md Showkat Hossain Chy

Md. Kamal Uddin

Samantha Perry

Cameron Brown

L. Carrington Goodrich

DOI：10.1016/j.jss.2025.112334

原文链接

原文求助

摘要

En 中文

Background: Understanding dependencies within microservices is essential for maintaining and evolving scalable and efficient software architectures. Dependencies influence how changes in one microservice might propagate to other microservices. With the decentralized nature of microservices, these dependencies might not be explicit to developers and lead to unique challenges in modern software development environments. Objective: The objective of this study is to synthesize existing literature on microservice dependencies, identify the types of dependencies, and examine the strategies employed to manage and analyze these relationships. This effort aims to elucidate how dependencies affect microservice systems and to provide a comprehensive overview of dependency management within microservices. Method: We conducted a multivocal literature review, starting with an initial dataset of 1,733 papers from academic literature (white literature). This corpus was narrowed down through a rigorous filtering process to 45 key publications that address the identification, management, and impacts of dependencies in microservices. Additionally, we incorporated 926 articles from grey literature sources such as Google, Stack Overflow, and Stack Exchange, expanding the scope beyond traditional academic research. After the filtration process, 45 articles were fully synthesized to integrate practical insights and professional experiences into our review. Results: The review identifies several types of dependencies in microservice systems and synthesizes this information into a unified dependency taxonomy. This review highlights a range of approaches to dependency management, revealing a significant gap in systematic catering approaches to generate taxonomies for dependencies and the need for integrated management tools. The findings underscore the fragmented nature of existing dependency management practices and the potential for more holistic approaches. Conclusion: This study provides valuable insights for researchers and practitioners, outlining effective strategies and pointing out areas needing improvement in dependency management. By offering a structured overview of the topic, the study serves as a roadmap for future research and development efforts to enhance the robustness and maintainability of microservices.

Keyword:

Microservice dependencies

Dependency taxonomy

Software maintenance

Distributed systems

Performance

Root cause analysis

Anomaly detection