A Switch Migration-based Congestion Control Mechanism for Multi-domain SDNs

Articles in Press

Document Type : Original Article

Authors

1 Islamic Azad University , Maybod Branch

2 Islamic azad university, taft branch

3 Department of Computer Engineering, Maybod branch, Islamic Azad University, Maybod, Iran

4 Islamic Azad University, Maybod branch

Abstract

With the continuous advancement of Software-Defined Networks (SDNs), the adoption of a distributed control plane architecture has become increasingly necessary. One of the primary challenges in these networks is the variable load on the controllers where high loads can lead to congestion. Such congestion can significantly degrade network efficiency. Although previous studies have attempted to address this issue, they have largely failed to effectively manage load exchange between the control plane and the data plane. This paper proposes a migration-based congestion control mechanism for multi-domain SDNs. In this approach, when a controller experiences high load and congestion, selected switches are migrated from the overloaded controller to one with a lower load. If the migration risks congesting the new controller, the mechanism swaps switches between controllers with minimal migrations, drawing inspiration from the Kadane algorithm to prevent congestion elsewhere. The proposed mechanism was evaluated using the D-ITG and IPerf tools with the RYU controller, demonstrating improved system performance. Simulation results show that the mechanism outperforms the baseline approach, increasing average network throughput by approximately 10%, while reducing average delay and jitter by about 30% and 25%, respectively. Furthermore, a comparison between the proposed method and the OptiGSM method reveals that the proposed method offers superior throughput and lower delay, although the OptiGSM method exhibits less jitter than the proposed method

Keywords

Main Subjects

Tabriz Journal of Electrical Engineering

Articles in Press, Corrected Proof
Available Online from 08 June 2025

Files

Share

How to cite

Statistics