Single servers that often go down due to overload or high workload, this can be handled by the Cluster Server Load Balancing scheme, but the performance of this scheme needs to be measured. This research purpose to compare the performance of two algorithms, namely the Round Robin and Least Connection algorithms in implementing the Apache load balancer in Docker container-based virtualization. The performance test used is load testing using the Apache JMeter tool. The parameters measured are throughput and response time. This research uses the SIDLC (System Infrastructure Development Life Cycle) method. The research results show that the Apache load balancer with the least connection algorithm is superior because it can handle 300,000 sample requests with a throughput value of 2120.9 requests/seconds and a response time of 238 milliseconds. The Apache load balancer with least connection can handle http-requests in 141 seconds, this is 9 seconds faster than the round robin algorithm which is able to handle 300,000 sample requests in 150 seconds, then the error value for the least connection algorithm is still below 0.05% during the highest test peak, 300,000 sample requests.
Copyrights © 2024