Swift Recovery: An Innovative Routing Approach for Multi-Failure Protection in Software-Defined Networks

Abstract

This paper proposes an innovative approach to routing for multi-failure protection in Software-Defined Networks (SDN). The goal is to minimize the number of flow configurations, thereby easing the burden on network switches and enhancing overall network performance. The approach categorizes the flow entries into permanent and recovery flows and leverages both an Integer Linear Programming (ILP) method and a heuristic method to generate multiple routing trees. These trees are strategically combined to reroute traffic during link failures, providing robust network resilience. Extensive evaluations across various network topologies, ranging from small to large-scale, revealed that our approach significantly outperforms the conventional Shortest Path (SP) method. Simulation results demonstrate a remarkable reduction in the number of flow entries required for initial connection configuration by 58.83% and for routing during multiple link failures by up to 95.44%, underscoring the effectiveness of the proposed method in managing multiple link failures in SDN. To demonstrate its practical application, the proposed method was implemented on the Ryu controller and integrated into a Pica8 switch, proving its effectiveness in providing rapid recovery without requiring additional hardware.