Comparative Analysis of Shortest Path and Load Balancing Algorithms in Software-Defined Networking Using Ryu Controller and Mininet

Abstract

Software-Defined Networking (SDN) enables programmable network management through the separation of the control plane and the data plane, making routing algorithm selection a critical factor in traffic distribution and network performance. The shortest path algorithm is commonly used as a baseline due to its deterministic behavior, but it may lead to load imbalance in topologies with multiple available paths. This study aims to compare routing behavior and network throughput between shortest path and Round Robin–based load balancing algorithms in an SDN environment. An experimental approach was employed by implementing both algorithms on the same topology and configuration using the Ryu controller and Mininet. Evaluation was conducted using observational and descriptive analysis through routing path inspection, OpenFlow flow rule analysis, and throughput measurement using iperf, a widely used network performance measurement tool that generates traffic to evaluate bandwidth between end hosts. The experimental results show that the shortest path algorithm consistently utilizes a single dominant path with an aggregate throughput of 6.46 Gbit/s, while the load balancing algorithm distributes traffic across multiple active paths and achieves a higher aggregate throughput of 9.65 Gbit/s. These findings indicate that load balancing provides better bandwidth utilization than shortest path under the tested experimental conditions.

Keywords

Software-Defined Networking; Routing Algorithms; Shortest Path; Load Balancing; Ryu Controller; Mininet