Comparação entre arquiteturas de software monolítica e de microsserviços

Abstract

This study investigates the persistent debate between monolithic and microservices software architectures, focusing on the transition from theoretical analysis to empirical validation. The main objective is to quantify the trade-offs associated with each approach in a controlled experimental environment. To this end, an experimental study was conducted, consisting of the implementation of functionally identical prototypes of a CRUD system—one with a monolithic architecture and the other with microservices—using the WEB JavaScript ecosystem with Node.js, Express, and Docker. The comparative analysis was performed based on a set of objective metrics, including API response time, CPU and memory usage, and the complexity of the deployment process. The collected data reveal that the monolithic architecture excels in lower latency and infrastructure simplicity. Conversely, the costs of the microservices architecture were quantified through a 147% higher latency in the orchestration operation, a 271% greater memory consumption at rest, and a complexity that required more than three times the number of configuration artifacts. As an academic contribution, this work presents a practical and reproducible case study that serves as a decision-support tool for software architects and development teams, grounding a strategic choice in quantitative and qualitative evidence regarding the inherent costs of each approach.