Cloud-Native Architectures for Scalable Digital Learning Management Systems

Abstract

Learning Management Systems (LMS) have evolved from monolithic on-premises installations to cloud-hosted platforms, but many institutional deployments still operate on architectures that cannot elastically scale to handle demand spikes -- semester start enrolments, live exam sessions, and concurrent video streaming peaks that can increase load 20-50x within minutes. Cloud-native architectural patterns -- microservices, containerisation, serverless functions, and event-driven messaging -- enable LMS platforms to scale elastically, deploy continuously, and recover from failures without downtime. This paper proposes the Cloud-Native LMS Architecture Framework (CNLAF), a systematic evaluation of five cloud-native architecture patterns -- monolithic cloud migration, microservices decomposition, serverless-first, event-driven microservices, and hybrid edge-cloud -- for digital LMS deployments across six performance and operational benchmarks. CNLAF is evaluated through load testing at 10K-500K concurrent users and real deployment measurement at three institutional LMS platforms over 12 months. Key results: event-driven microservices achieves the highest LMS Architecture Score (LMSAS = 0.924) with 99.98% availability and 124 ms p99 response time at 500K concurrent users; serverless-first achieves the lowest operational cost (0.42 USD per 1,000 active user-hours); hybrid edge-cloud reduces video streaming latency by 68% for geographically distributed learners.