Adaptive Failure Recovery And Path Optimization For Multi-Hop Software-Defined Exchange Point Networks

Mohamed, Abdijalil Abdullahi (2025) Adaptive Failure Recovery And Path Optimization For Multi-Hop Software-Defined Exchange Point Networks. PhD thesis, Universiti Sains Malaysia.

[img] PDF
Download (642kB)

Abstract

This thesis presents the Enhanced Link Failure Rerouting (ELFR) mechanism, developed to address challenges in failure recovery within multi-hop Software-Defined Exchange Point (SDX) networks. SDX, an evolution of traditional Internet Exchange Points (IXPs), uses Software-Defined Networking (SDN) to enable programmable traffic control. However, multi-hop SDX deployments still face significant limitations, including high packet processing delays, slow path computation, and excessive memory consumption during failures. To tackle these issues, ELFR was designed with two integrated modules: the Packet Processing Module, which allows fast in-switch classification and recovery using P4-defined headers, and the Path Computation Module, which proactively calculates neighbor-based backup paths to lessen controller overhead. The architecture was implemented with Mininet, BMv2, P4Runtime, and the POX controller, and tested across four realistic failure scenarios. Results show that ELFR outperforms the Flow-Aware Pro-Reactive (FAPR) mechanism by reducing Recovery Time (RT) by 46.97%, Computation Time (CT) by 34.62%, and Computational Overhead (CO) by 44.44%, while decreasing Memory Usage (M) by 37.5%. Additionally, it minimizes Packet Processing Delay (PPD) by 28% and Packet Loss Ratio (PLR) by 63.16%. These findings confirm ELFR as an efficient, scalable, and resilient solution for recovering from link failures in programmable SDX infrastructures.

Item Type: Thesis (PhD)
Subjects: T Technology > T Technology (General) > T1-995 Technology(General)
Divisions: Pusat IPv6 Termaju Negara (National Advanced IPv6 Centre of Excellence NAv6) > Thesis
Depositing User: Mr Aizat Asmawi Abdul Rahim
Date Deposited: 09 Sep 2026 07:39
Last Modified: 09 Sep 2026 07:39
URI: http://eprints.usm.my/id/eprint/64907

Actions (login required)

View Item View Item
USM Repository is powered by EPrints 3 which is developed by the School of Electronics and Computer Science at the University of Southampton. More information and software credits.
Share