<![CDATA[Backend constitutes a critical component of digital infrastructure, responsible for processing business logic, managing data, and facilitating communication between software systems. APIs serve as the interface that enables software interaction and plays a pivotal role in backend operations. This study investigates the performance of three API architectures: RESTful, GraphQL, and gRPC. The experimental approach involves the implementation of Load Testing and Stress Testing to assess the performance of these architectures. The experiment utilizes a dedicated server and client hardware to simulate real- world conditions, with parameters such as CPU usage, memory usage, response time, load time, latency, success rate, and failure rate evaluated using a dataset comprising 1,000 rows of student- related records. Result show that RESTful achieves the highest total request but exhibit greater resource consumption and a higher failure rate. GraphQL demonstrated better CPU and memory efficiency with strong stability, though it has higher latency and slower response times. gRPC strikes a balance with a moderate latency and resource usage, albeit with slightly higher memory consumption under stress. By presenting a comprehensive analysis of each API architecture, this study contributes a comprehensive performance analysis under practical testing scenarios giving developers and system architect with data-driven guidance for selecting API architecture to their application needs. RESTful is well suited for high-throughput scenarios with less critical operations, GraphQL excels in resource efficiency and stability, and gRPC offers balanced performance across diverse workloads.]]>
Copyrights © 2025
