<![CDATA[The increasing number of users and web service requests demands a system capable of dynamically adjusting resource capacity to maintain performance and service availability. One solution is the use of load balancing combined with automatic scaling (auto scaling) in the OpenStack environment. This study analyzes the performance of a load balancer system using an auto scaling approach on the Apache2 web server in OpenStack. The methodology employed is the Network Development Life Cycle (NDLC), which includes the Analysis phase to identify system requirements and resources. The Design phase covers the design of network topology and testing scenarios, while the Simulation phase involves the installation and configuration of OpenStack. Testing was conducted using three scenarios: CPU load, HTTP requests, and memory load, each performed five times. Monitoring was carried out using Gnocchi services. The results indicate that the system effectively responds to increasing load while maintaining web service stability. CPU load testing showed an average scale-up time of 583 seconds and scale-down time of 213 seconds. HTTP request testing with 1,000,000 requests resulted in an average scale-up of 543 seconds and scale-down of 297 seconds, while memory load testing produced a scale-up of 561 seconds and scale-down of 81 seconds. In conclusion, the implementation of load balancing with auto scaling is considered effective, although Gnocchi metrics are aggregate and do not fully represent real-time conditions.]]>
Copyrights © 2026
