Performance evaluation of ryu, opendaylight and floodlight controllers in diverse software-defined networking topologies

Software-Defined Networking (SDN) has been introduced as a new approach to networking for designing and managing computer networks. SDN architecture decouples control and data planes, enabling programmable network management. The significant component is the Co...

Full description

Bibliographic Details
Main Authors: Hui, Diong, Hoh, Wei Siang, Ong, Bi Lynn, Zhu, XinPing, Yoon, Si-Kee
Format: Article
Language:English
Published: Akademia Baru 2025
Subjects:
Online Access:https://umpir.ump.edu.my/id/eprint/45281/
Description
Summary:Software-Defined Networking (SDN) has been introduced as a new approach to networking for designing and managing computer networks. SDN architecture decouples control and data planes, enabling programmable network management. The significant component is the Controller, which is responsible for managing and distributing information to all network devices. While prior studies evaluate SDN controllers, comprehensive comparisons across multiple topologies and performance metrics remain limited. This paper addresses this gap by analysingand comparing the performance of the Ryu, OpenDayLightandFloodlight controllers under single, linearandtree network topologies using Mininet. The performance parameters considered include round-trip time (RTT), throughput, jitterandpacket delivery ratio. The results indicate that in linear and tree topologies, Ryu exhibited the lowest RTT, highest bandwidthandlowest jitter compared to the ODL and Floodlight controllers. Conversely, in the single topology, Floodlight demonstrated the lowest jitter, while Ryu maintained the lowest RTT and highest bandwidth. In conclusion, Ryu is often preferred in scenarios requiring lower latency and higher throughput, particularly in more complex network topologies. However, Floodlight maystill be advantageous in simpler topologies where minimizing jitter is critical. This comprehensive evaluation not only aids in controller selection but also informs future enhancements in SDN architecture and performance benchmarking methodologies, whichwill be implemented in modern data centres.