<![CDATA[Digital transformation in Indonesia’s healthcare sector demands a Hospital Information System (SIMRS) that is scalable, reliable, and easy to integrate. Most existing SIMRS still rely on a monolithic architecture, making them vulnerable to single points of failure, hard to scale per module, and prone to downtime during updates. This study aims to examine strategies for migrating monolithic architectures to microservices and to evaluate the performance of a Docker- and Docker Swarm-based deployment on a SIMRS prototype consisting of twelve functional modules. The method combines a Systematic Literature Review on recent decomposition patterns with an experimental study on a two-node cluster running fourteen stacks and thirty-five services for a total of seventy-six active containers. Measurements show that the average HTTP response time ranges from 0.6 to 5.1 milliseconds for frontends (with a single billing-frontend outlier of 82.4 ms attributable to memory saturation) and from 0.8 to 3.0 milliseconds for backends, with per-container CPU utilization mostly below two percent and stable memory consumption within allocated limits. Docker Swarm proves effective in supporting high-availability replication, rolling updates, and automatic self-healing under normal operating conditions; one observed rolling-update failure (billing-backend dropping to 0/2 replicas until manual intervention) shows that reliable image-level health checks are a prerequisite for the promised zero-downtime behavior. The study concludes that microservices with Docker Swarm are a feasible adoption strategy for SIMRS in Indonesia, with manageable operational trade-offs.]]>
Copyrights © 2026
