Organizational Cohesion in Microservice Architectures: A Multi-Project Empirical Study
2026-06-15 • Software Engineering
Software Engineering
AI summaryⓘ
The authors study how teams of developers work on microservice software and how their activities match the software's modular structure. They create a new way to measure how focused and balanced teams are when working on individual microservices, called Pairwise Team Cohesion (PTC). By examining several open-source projects, including Spinnaker, they find that PTC is only loosely related to how much different teams work across services, suggesting that teamwork and code dependencies are separate organizational patterns. Their work helps understand how the team structure relates to the software design in microservice systems.
Microservice architectureSoftware modularityOrganizational cohesionPairwise Team Cohesion (PTC)Service couplingDeveloper activitySpinnakerOpen-source softwareSocio-technical structureLongitudinal case study
Authors
Xiaozhou Li, Andrea Janes
Abstract
The widespread adoption of microservice architectures has introduced new challenges in aligning software modularity with the structure of development organizations. Although prior research has extensively examined technical properties such as service coupling and dependency structures, comparatively little attention has been paid to how contributor activity reflects or diverges from service boundaries. In this paper, we introduce the notion of organizational cohesion in microservice ecosystems and propose a quantitative approach to measure it. Building on the Sensitive Class Cohesion Metric (SCOM), we define Pairwise Team Cohesion (PTC), a metric that captures the balance and focus of developer contributions within individual microservices. We analyze the evolution of organizational cohesion using a longitudinal case study of the Spinnaker microservice platform and replicate the analysis across six additional open-source microservice systems. Our results reveal systematic differences between core and peripheral services and show that PTC and Average Organizational Coupling (AOC) exhibit only a weak correlation across projects. This finding shows that team cohesion and cross-service developer activity suggest distinct and weakly associated organizational dynamics. By extending the "high cohesion, low coupling" principle to the organizational level, our study provides a quantitative perspective for assessing the socio-technical structure of microservice development.