Network Coding-based Resilient Routing for Maintaining Data Security and Availability in Software-Defined Networks

Haoran Ni, Zehua Guo*, Changlin Li, Songshi Dou, Chao Yao, Thar Baker

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)

Abstract

Software-Defined Networking (SDN) improves network performance by flexible traffic control. Data security and data availability are two main concerns for designing a resilient routing algorithm in SDN. Existing algorithms such as the MPT algorithm consider joint data security and availability, but they cannot make a good trade-off. In this paper, we propose a Network Coding-based Resilient Routing algorithm named NCRR to jointly achieve data security and availability. NCRR is a heuristic algorithm that computes routing decisions in three scenarios based on the number of disjoint paths. Specifically, the scenario of three disjoint paths is enough to ensure joint security and availability for a flow when three or more disjoint paths can be used for forwarding this flow. However, due to topological diversity, we cannot always find three disjoint paths for each flow. Thus, the scenarios of two disjoint paths and one path are used to ensure joint security and availability if only two disjoint paths and one path can be used. To evaluate the performance of NCRR, simulations have been conducted using two real-world network topologies. Simulation results show that NCRR improves the joint data security and availability performance by approximately 6.23% on AttMpls topology and 21.34% on Cernet topology, compared with existing MPT.

Original languageEnglish
Article number103372
JournalJournal of Network and Computer Applications
Volume205
DOIs
Publication statusPublished - Sept 2022

Keywords

  • Data availability
  • Data security
  • Routing
  • Software-Defined Networking

Fingerprint

Dive into the research topics of 'Network Coding-based Resilient Routing for Maintaining Data Security and Availability in Software-Defined Networks'. Together they form a unique fingerprint.

Cite this