<![CDATA[Single database architectures often experience performance degradation as user load increases because all read and write operations are processed centrally, causing bottlenecks and limiting the scalability of web systems. To address this issue, this study implements database replication using a MySQL Master–Slave architecture combined with the separation of read and write operations in a containerized Laravel-based web system. A quantitative experimental method was applied by comparing the Single Database and Master–Slave Replication configurations implemented in a Docker environment on Windows Subsystem for Linux 2 (WSL2). Performance evaluation was conducted using Apache JMeter at three workload levels low (10 users), medium (50 users), and high (100 users) based on response time, throughput analysis, latency analysis, and error rate. The results show that the Master–Slave configuration provides more stable performance at medium to high workloads, with latency between 0.20 and 0.26 seconds and an error rate of 0% in all test scenarios. Additionally, the highest throughput of 2.01 requests per second was achieved at medium workload, indicating effective read-write separation and better workload distribution through database replication.]]>
Copyrights © 2026
