Evaluation of the impacts of decomposing a monolithic application into microservices: A case study

TL;DR

This case study evaluates the effects of transitioning a financial industry application from a monolithic to a microservices architecture, showing improvements in modularity, CPU, and memory efficiency.

Contribution

It provides empirical evidence on the impacts of microservices decomposition on a real-world application, highlighting benefits and challenges.

Findings

Improved modularity metrics in microservices architecture

Reduced CPU and memory consumption after decomposition

Insights into challenges and future research directions

Abstract

Monolithic applications are being decomposed to a microservices architecture, aiming to improve maintainability, performance and modularization. Although such decompositions have now been widely carried out in the industry, little is reported in the literature about the impacts of these decompositions. This work, therefore, reports a case study carried out to investigate the impacts of the decomposition of a real industry application for microservices architecture. The target application of the study refers to a raid operation, performed by a financial institution, which was extracted from a monolithic application to an application based on microservices. In particular, metrics were applied to monolithic and microservice-based applications, aiming to quantify coupling, cohesion, CPU consumption and memory consumption. The results obtained show that the microservices architecture generated better results for the software modularity metrics, in addition to lower memory and CPU consumption. Finally, this work brings insight and points out challenges and future directions of research that need to be explored by academia and industry.