This study aims to compare the performance of a monolith database architecture (a single server for read and write operations) with a distributed database architecture that implements read–write separation using a replication mechanism in MySQL. The focus of the test was how the throughput, latency, and resource utilization (CPU, memory, and I/O) of the two architectures differed when given the same workload scenario, namely heavy read, heavy write, and balanced read write. The results of the study are expected to provide practical recommendations for system developers and managers in choosing a database architecture that is more suitable for applications with high transaction needs, especially in MySQL-based environments that are commonly used in the industry.
Copyrights © 2025