Skripsi
SISTEM LOAD BALANCING SERVER SMART FARMING BERBASIS DOCKER SWARM MENGGUNAKAN SINGLE BOARD COMPUTER
Smart farming requires a reliable server infrastructure to handle real-time data requests. The use of a single server may cause overload and reduce service performance. This study aims to implement a Docker Swarm-based load balancing system for smart farming servers using Single Board Computers in the form of Raspberry Pi. The system architecture consists of one manager node and two worker nodes with NGINX as a load balancer using the least connection algorithm. Performance testing was conducted using stress tests with traffic variations from 1,000 to 15,000 requests. The results show that load balancing reduces the average response time from 351.75 ms to 274.10 ms and increases throughput from 584.45 req/s to 904.22 req/s, while maintaining service availability during node failures.