Optimization of the latency in networks SDN

Unlike traditional networks, software-defined networking (SDN) are characterized by a physical separation of the control and the transfer plan. Thus, a centralized controller communicates the functions of control plan to each device via the OpenFlow Protocol whenever he is asked or that it deems appropriate. This impact strongly the latency time which is important for new services or multimedia applications. In order to optimize the time of transmission in network data with SDN, the proactive approach based on the algorithm back - pressure is usually oered. However, we note that the proactive approach while reducing strongly this time, does not account settings such as the failure of a node part of the way to transfer or the breaking of a bond that greatly increases the latency time. In this document, we will propose a joint routing approach based on proactive and reactive routing. This in order to optimize the routing functions by simply placing the traffic where capacity allows, in order to avoid congestion of highly stressed parts of the network taking into account the failures and significantly reduce the time latency. Simulation results show that our proposal allows a considerable reduction of latency even when there are failures in the network.