Improving link failure recovery and congestion control in SDNs

Software-Defined Networking is considered as a prominent architecture in the field of networking. The separation of control plane with data plane provides an efficient way to implement and manage real-world networks. In recent years, there has been a major boost in network traffic due to the advancement in IoT and cloud technologies. The main drive behind the SDNs is primarily due to the fact that existing legacy networks do not have the capacity to manage such large volumes of traffic in coming future. One of the major concern related to the implementation of SDN networks is the recovery time in case of link failure and congestion. This is known to be the main bottleneck for the scalability of pure SDN networks. The existing method in use requires the SDN devices to contact the centralized controller to act during such events. However, in real-world networks consisting of many devices, such a solution increases the latency immensely thus reducing the efficiency of the network. In here, we propose a method to improve the recovery time for link failure and congestion events. The method consists of controller modules that proactively install rules before the occurrence of such events. This will provide the ability to these devices to handle such issues at the data plane without the intervention of controller. The rules are later pushed to SDN devices at regular intervals using different algorithms and RESTful methods. The proposed solution largely improves the efficiency of these networks by reducing their overall latency.