Research On SDN-based Traffic Classification Method And Scheduling

In recent years,with the rapid development of the Internet,complex and diverse applications and network traffic have been generated,and traditional network architectures and equipment have been unable to meet the increasing data traffic in limited bandwidth.Software-defined networking is an architecture that is considered to replace the current traditional network technology.SDN separates control and forwarding to simplify the network structure,improve network throughput and user service quality,and achieve fine-grained network management.The main content of this article is to classify traffic and use different scheduling algorithms for path planning,as follows:(1)Aiming at the problem of traffic classification in SDN network,in order to improve the efficiency of traffic classification and reduce the classification time,this paper proposes a network traffic classification model(Cosine similarity and decision tree classification model,CSDT)based on similarity and decision tree algorithm.To identify and classify traffic.This model takes advantage of the high similarity of adjacent data streams in the process of network traffic interaction,and uses similarity algorithms to preprocess network traffic to achieve the purpose of reducing classification time.The text uses the Moore data set publicly available in the field of network traffic classification for training and testing,and the results are compared with various machine learning algorithms on the Weka platform.The experimental results show that the model greatly reduces the classification time when the classification accuracy is good.Improve the efficiency of classification of network traffic.(2)Aiming at the routing optimization problem in SDN networks,in order to improve network throughput and user service quality,reduce network utilization and forwarding delay,this paper proposes a traffic scheduling model based on ?-Q-Learning algorithm.In the traffic scheduling model of this thesis,by judging the traffic service type,different forwarding paths are provided for the traffic of different service types.If it is background and interactive traffic with a lower Qo S level,the shortest path algorithm or ECMP algorithm is used for path planning;On the contrary,use ?-Q-Learning algorithm for path planning.The model experiment part simulates four traffic models for testing according to different online behaviors of users.The results show that the traffic scheduling model proposed in this paper based on ?-Q-Learning has good performance in average throughput,link utilization,bandwidth utilization,round-trip delay and overall bid loss rate.