Wireless Router Management And Abnormal Traffic Detection Based On SDN Network

With the rapid development of communication technology and the release of restrictions on mobile WLAN by the Ministry of Industry and Information Technology,wireless network equipment has become popular in large numbers.Meanwhile,network security is becoming more and more important.In the face of network security problems and limited network bandwidth,it is difficult to meet the needs of users.Large-scale deployment of traditional networks leads to complex network structures and difficult unified management and maintenance.Compared with the traditional distributed Network,software-defined network(SDN)has better advantages.It adopts a centralized control mode,which is similar to 4G network architecture,separates the data plane from the control plane,and has lower latency.Aiming at the architecture characteristics of SDN,this paper designs a management platform based on SDN,which is used to manage the flow table and bandwidth of network equipment.Aiming at the security problems of abnormal traffic detection in SDN,a feature selection algorithm and abnormal network traffic detection model based on machine learning are designed to detect whether there is abnormal traffic in the network.The main research work of this paper includes:(1)Study the characteristics of SDN network and SDN controller,design a set of wireless router network management platform based on SDN according to the characteristics of centralized management of SDN controller,realize the configuration management of wireless router flow table by using openflow and other protocols,complete the configuration function of network equipment flow table and metering meter.In this way,flow tables and metering tables can be configured in a centralized manner on the management system to implement user-defined forwarding policies and bandwidth control based on metering tables.(2)The abnormal traffic detection problem is transformed into a binary machine learning problem,and the feature selection algorithm and abnormal traffic detection algorithm based on machine learning are designed.In this paper,open source data set is used.Firstly,the improved feature selection XGBoost-FRA algorithm is used to select the best features,which overcomes the defect of single XGBoost for feature selection and realizes the fast selection of effective features.The open source data set is then used as the exception detection data set.Through the improved anomaly detection BOLight GBM model,the model has the advantages of higher accuracy and less detection time for abnormal traffic.(3)This paper constructs the experimental environment of SDN network,tests the SDN network management system,realizes the self-defined forwarding and bandwidth control,and achieves the desired effect.Then,the anomaly detection model based on SDN network is tested.The open source NSL-KDD data set is used as the input to train the anomaly detection classification model and realize the anomaly detection classification.The experimental results show that,compared with other algorithms,BOLight GBM model has better overall performance,and the validity of this abnormal traffic detection method has been verified.