Dynamic Scaling in e-commerce Platforms: Microservices for Latency, Compliance, and Resilience

Abstract

This article presents research on the use of Federated Reinforcement Learning (FRL) as an intelligent mechanism for enabling compliance‑aware microservices and dynamic scaling in multi-cloud environments. Modern regulations such as the General Data Protection Regulation (GDPR), the Health Insurance Portability and Accountability Act (HIPAA), and the Payment Card Industry – Data Security Standard (PCI‑DSS) impose strict constraints on how distributed systems handle sensitive data. Traditional approaches to autoscaling and microservices orchestration often require centralized data aggregation for performance optimization, creating conflicts with privacy requirements. FRL provides a privacy‑preserving alternative by enabling learning‑driven scaling and decision‑making across multiple cloud platforms without exposing raw data.

In this study, FRL is applied to enhance microservices‑based dynamic scaling, enabling systems to optimize latency, stability, and compliance simultaneously. Experimental results show that FRL‑driven microservices outperform traditional centralized scaling models in accuracy, responsiveness, and regulatory alignment. Furthermore, the comparison among centralized, federated, and hybrid learning models reveals that the hybrid FRL‑enhanced approach provides the best balance of speed, compliance, and performance for highly dynamic e‑commerce workloads.

Industries such as finance, healthcare, and e‑commerce where privacy, real‑time observability, and regulated processing are essential stand to benefit significantly from FRL‑enabled microservices architecture. The findings highlight FRL’s potential to become a core governance‑enhancing and performance‑optimizing component in future multi-cloud and cloud‑native systems.