Routing Optimization in SDN using Scalable Load Prediction

With the exponential growth of data traffic and the rapid development of smart devices, networks are becoming more and more heterogeneous and complex. Therefore, managing network resources with traditional routing features is no longer advised. More intelligence needs to be deployed. However, deploying intelligence in traditional networks seems to be hard to achieve since they are naturally distributed. The emerging of Software Defined Networking (SDN) will ease the introduction of intelligence in networks. In this vein, Deep Learning (DL) is considered the most promising concept for intelligence delivery.In this paper, we combine Deep learning (DL) with SDN in order to improve the performance of routing techniques efficiently. In this work, we analyse the algorithm denoted “Predicting of Future load-based routing (PFLR)” and we prove that it is not scalable in Very Large-Scale Networks(VLSN). Therefore, we suggest an enhancement of this algorithm to achieve scalability by predicting available future bandwidth on path-basis in spite of link-basis. Predicted values are obtained using a MultiLayer Perceptron (MLP) neural network and applied in the Dijkstra algorithm to find the optimal path according to a reciprocal metric. The proposed approach is denoted Scalable Predicting of future load-based routing (SPFLR). Experiments show that the proposed approach outperforms parallel ones by achieving significant load balancing through the network.