Traffic Classification and Comparative Analysis with Machine Learning Algorithms in Software Defined Networks

In computer networks, diverse applications generate network traffic with different characteristics. Network classification is significant to manage networks better, improve service quality and ensure security. Software-Defined Networks (SDN) provides flexible adaptable techniques for its programmable structure. SDN flows naturally exhibit particular characteristics of protocols. Therefore, it can be said that present opportunities in using machine learning. This study proposes a approach learning models SDN. this study, DNS, Telnet, Ping Voice were created on the Distributed Internet Traffic Generator (D-ITG) tool. Twelve-attributes representing these (the number packets transmitted, average transmission time, instantly transmitted packets, etc.) determined, over controller physical network, real-time data set was by collecting depending attributes. Later, performance k Nearest Neighbor (k-NN), Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), Decision Tree (DT) Naive Bayes (NB) tested set. When k-NN model set, accuracy obtained as maximum 99.4%. Therefore has been determined giving highest lowest cost flow attributes