A Resilient Response Method for DDoS Attacks Based on Deep Learning in NFV

The widespread distributed denial of service (DDoS) attacks has become a major threat to network security. The vigorous development of network function virtualization (NFV) technology has provided a new idea to deal with DDoS attacks. In view of the weak defense capability of the current DDoS attack response methods in NFV, this paper proposes a resilient response method for DDoS attacks based on deep learning. First, a two-stage traffic detection and cleaning device based on information entropy and convolutional neural network is designed and a distributed deployment method of traffic detection and cleaning device is determined to improve the accuracy of traffic detection. Second, a resilient mechanism of the traffic detection and cleaning device is proposed to realize the rapid scalability of resources on demand and improve the ability of networks to deal with DDoS attacks. Experimental results show that the resilient response method proposed in this paper improves the detection accuracy and resource utilization, and realizes the flexible deployment and resilient scalability of the traffic detection and cleaning device. Compared with other similar methods, the performance of the method proposed in this paper is superior.