SDN DDoS Defence: A Novel Combinatoral Optimization via Meta-Heuristic Approach with Deep Convolutional Neural Network

Sukhvinder Singh and S.K.V. Jayakumar

Computer Network has become very complex since the last few decades. The ability of self-controlling and managing is the essential necessity of today's world of network. The functioning of the traditional network is entirely depending on the hardware. Any change in the network leads to the replacement or increase in the hardware. IT industries are also looking forward to the future network with their own expectation such a network which can evolve, create and permits changes with unlimited possibilities. Software Defined Network (SDN) redefines the network architecture by realizing the physical to a logical link of network hardware and software. It has the ability to change and create the future together. SDN architecture offers immense opportunity for future network. Further, all the technologies will have some shortfalls which need to be patched up at the initial stage. Since last few years the Distributed Denial of Service attack has become very common, and if the controller is found suspectable then the attackers have full chances to take control over the network and its resources. Hence, to detect DDoS flow at early stage a robust security mechanism is needed. Proposed Meta-Heuristic Based Optimization approach for DDoS attack detection has two-parts; feature extraction and classification. The Particle Swarm Optimization; a meta-heuristic algorithm is used for feature extraction and Deep Convolution Neural Network will classify the input features and detect the DDoS attack. The experimental outcome proves that the DDoS attack detection using Meta-Heuristic Based Optimized Deep CNN outperforms as compared to other schemes.

Volume 11 | Issue 6

Pages: 95-102