Abstract The SDN has increased its focus, and the notion of network control has offered an efficient network-oriented DDoS protection in addition to many DDoS assault methods. More information about the network might be influenced by the centralized SDN controller, and SDN framework helps in identifying DDoS assaults using various methods. The simulation dataset for this work was generated by constructing SDN on the Mininet emulator. To construct the dataset and train the deep learning algorithm, the unique features are logged into a csv file. Further, detection is done using Optimized Deep Max out classifier. In addition, the weights of Deep Max out classifier are chosen via Sine Map Insisted CA (SMI-CA) model. If any attack is found, Bait oriented mitigation is made for relieving from attacks. As last step, analysis is done to portray the effectiveness of adopted model. The model used in the paper is further evaluated using the newly released dataset CICDDoS2019 along with the simulation dataset. Result shows that the Deep Maxout classifier has a very low false alarm rate and can classify traffic with the greatest testing accuracy of 96.5% for the CICDoS2019 dataset and 95.1% for the simulation dataset.
Field : Mühendislik
Journal Type : Uluslararası
Relevant Articles | Author | # |
---|
Article | Author | # |
---|