Research On Intelligent Security Detection Method Based On SDN

Software Defined Network(SDN)is a form of Internet architecture that has emerged in recent years to simplify traditional network architectures with its scalability,dynamism and programmability.The architecture realizes centralized management by separating the data forwarding plane and control plane of the network.With the continuous progress of intelligent technology,the establishment of the electric Internet of Things is no longer limited to the network isolation of a single LAN industrial environment,but gradually extended to the Internet.In order to deal with the high network transmission pressure in the power Internet of Things,SDN technology can be used to realize the separation of the data plane and control plane,as well as network virtualization,implement efficient network transmission strategy,and establish efficient SDN architecture under the power Internet of Things.However,due to the characteristic of centralized management of SDN,SDN-based electric iot is more vulnerable to attacks in network transmission than traditional electric iot.Once attacked,the overall paralysis of electric iot may be caused.Based on this,this paper studies the intelligent security detection method in the SDN environment of the electric Internet of Things.This paper studies the SDN related technologies and the theory and architecture of the power Internet of things.Based on the analysis of the network security risk of the power Internet of things,the SDN-based network security detection model is studied in the power Internet of things.Distributed denial of Service attacks,often referred to simply as DDoS attacks,are the most aggressive of all attacks,and these attacks generate many packets that eventually crash the attack target system.In this paper,a hybrid neural network architecture called DDoSTC is designed,which combines the efficient and scalable Transformer and Convolutional Neural Network(CNN)to detect DDoS attacks on SDN,and is tested on the latest dataset CICDDoS2019.For better verification,the dataset was divided and compared with the latest deep learning detection algorithm currently applied in the field of DDoS attack detection.Experimental results show that the average AUC of DDoSTC is 2.52%higher than that of the current optimal model,and DDoSTC is more successful than the current optimal model in terms of average precision,average recall and F1 score.In this paper,we demonstrate the feasibility and superiority of our proposed detection model in terms of accuracy,precision and other metrics by training and validating a real electricity dataset and a dataset collected from an SDN environment of the electric Internet of Things endpoints.We then conduct simulated experiments on DDoS attacks,and our security detection mitigation mechanism has been validated to effectively identify and mitigate DDoS attacks in SDN-based electric Internet of Things environments,thus ensuring the normal operation of the network.