The Research Of Self-healing For OpenFlow Networks

Development of mobile internet and Virtualization Technology put forward higher requirements on network self-healing that increases the reliability of network data transmission. OpenFlow networks have been proposed as an innovative split architecture that decouples control plane from the forwarding data plane, which greatly promote the development of the research of self-healing.The ability to fast failover from failures has always been a crucial goal of network self-healing in OpenFlow architecture. In this paper, we have researched the selfhealing scheme for survivable OpenFlow networks according to failure detection and transfer path failover phase. The self-healing scheme mainly proposed two key technologies to ensure the fast failover: Local Reverse Back algorithm and Node Classification-based Fast Failover Mechanism, which to ensure fast failover from network link failures. Local Reverse Back algorithm implements efficient link failure detection and notification without extra detection mechanism and it selectively reverse traffic back to upstream node to make a comprehensive failover decision. Based on categorizing switches of data plane into a set of types, Node Classification-based Fast Failover Mechanism establishes an efficient autonomic coordination mechanism between restoration and protection, which to choose an optimal fast failover scheme. To demonstrate the practical performance of the proposed scheme, we have implemented NCFF in our OpenFlow testbed. Compared with related research results, the numerical results verify that our self-healing scheme enables faster reaction time and high survivability in various topologies.