Improving MPTCP Performance by Detouring Network Congestion using Software-Defined Networks

Bong Hwan Oh*, Yikun Li

*Corresponding author for this work

Research output: Chapter in Book or Report/Conference proceedingConference Proceedingpeer-review

Abstract

Multipath TCP (MPTCP) is a protocol that operates at the transport layer and is designed to facilitate multipath transmission, which can significantly improve network performance and more efficiently utilize network resources. However, the original MPTCP has been known to suffer from throughput degradation if one of the subflows experiences congestion, resulting in underutilization of network capacity when compared to a single TCP. In this research, we propose a novel approach to detour a congestion situation for MPTCP subflows using Software-Defined Networking (SDN), which can help to achieve more stable MPTCP performance. Our experimental results demonstrate that our proposed scheme helps minimize the impact of network congestion and enables MPTCP to perform efficient packet transmission.

Original languageEnglish
Title of host publicationAI2A 2024, Conference Proceedings - 2024 4th International Conference on Artificial Intelligence, Automation and Algorithms
PublisherAssociation for Computing Machinery
Pages150-153
Number of pages4
ISBN (Electronic)9798400717840
DOIs
Publication statusPublished - 4 Dec 2024
Event4th International Conference on Artificial Intelligence, Automation and Algorithms, AI2A 2024 - Shanghai, China
Duration: 27 Sept 202429 Sept 2024

Publication series

NameACM International Conference Proceeding Series

Conference

Conference4th International Conference on Artificial Intelligence, Automation and Algorithms, AI2A 2024
Country/TerritoryChina
CityShanghai
Period27/09/2429/09/24

Keywords

  • Multipath TCP
  • Network Traffic Engineering
  • Software-Defined Networks

Fingerprint

Dive into the research topics of 'Improving MPTCP Performance by Detouring Network Congestion using Software-Defined Networks'. Together they form a unique fingerprint.

Cite this