No abstract

No abstract

Software defined network (SDN) enables network programmability and provides fine grained control for managing the complex network infrastructure. With the centralized nature of SDN it poses the requirement for failure management at the data plane, control plane and at the centralized controller side. The self-healing attribute of the autonomic network can be combined with SDN to develop a software defined self-healing resilient network. The aim of this paper is to propose a self-healing SDN framework which can optimize the recovery by applying autonomic principles. The proposed work includes a rapid recovery (RR) mechanism to perform an immediate link recovery at the switch level without overburdening the controller. Additionally, it reduces the memory requirement of the switch for storing the backup path flow rules by aggregating all the disrupted flows. We presented the analytical model for calculating the failure recovery time and the backup flow rules required for recovery. Based on the analytical model, RR scheme reduces the total number of backup flow rules for all the disrupted flows of the failed link to a single flow rule.

One of the key technical challenges in 5G is to ensure a high level of end-to-end service availability and reliability. Software-defined networking leveraging on a centralized network-forwarding control and management paradigm is expected to fulfil such requirements and even enhance their capabilities, compared to the traditional reliability mechanisms in the legacy network. This paper presents an enhanced forwarding table design for scalable and efficient failover support on multilink failures in OpenFlow networks. The proposed forwarding table is designed with flow grouping and flow aggregation methods to achieve faster link failover and compress the flow rules for alternate paths against multi-link failures. To validate the proposed forwarding table design, we evaluate the performance in the following points: (1) how much resilience can be provided, (2) how many forwarding rules can be reduced to configure the alternate paths against multi-link failures and (3) how rapidly the failover can be achieved. Emulation results confirm that the proposed design reduces the number of disrupted flows and required forwarding rules to configure the alternate paths. We could also identify that it can achieve the recovery in 2-3 ms.

No abstract