Análisis comparativo de patrones de diseño en arquitecturas monolíticas y de microservicios: Impacto en el rendimiento, escalabilidad y mantenibilidad

Abstract

Abstract: This study compares the impact of design patterns on monolithic and microservices architectures by evaluating performance, scalability, and maintainability under the ISO/IEC 25010 quality standard. The experimental design consisted of two groups: monolithic systems implemented with the Django framework, and their microservices counterparts developed using FastAPI and deployed with Docker. Each architecture was implemented both with and without design patterns, resulting in four functional prototypes. Based on these implementations, fourteen testing scenarios were configured in Apache JMeter to evaluate performance and scalability, while maintainability was assessed using SonarQube. A total of 42 execution runs processed 4,179,278 HTTP requests, enabling a comprehensive comparative analysis. The results indicate that monolith_patterns achieved the highest overall score (0.9149), with a baseline latency of 8.5 ms, zero technical debt, and the maximum scalability score. In contrast, micro_patterns led in raw performance (0.9133) but required optimized infrastructure. The incorporation of design patterns improved overall performance by 151%. The architectural decision threshold was identified between 200 and 500 concurrent users: below this range, the monolith with GoF patterns performs better; above it, properly configured microservices outperform in throughput.