DDOS attack detection method for SDN by using deep neutral network

• Main Authors: Reddy, Kumbala Pradeep, Devi, G. Naga Rama, Prakash, S. Wilson, Srinath, B.
• Format: Conference Proceeding
• Language: English
• Subjects: Artificial neural networks; Comparative studies; Cybersecurity; Damage; Denial of service attacks; Machine learning; Malware; Neural networks; Optimization; Security systems; Software-defined networking; Task analysis
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0118429

The controlling action of network through software logic in innovative architecture of Software Defined Network (SDN) in which optimization of traffic flow, scalability and simplicity are projected. But user damage amount increment is causes the malware in network. So, potential strength is required to network software for counter malware. One of the fast growing attacks is Distributed Denial of Services (DDoS) which are causes a great threat to internet. These DDoS attacks are detected efficiently by using machine learning (ML) methods. A DDoS attack detection method for SDN by using Deep Neural Network (DNN) model is analyzed in this paper. A number of DDoS attack detection and mitigation models are explained by many researchers but these DDoS attacks are improving their strength day to day and damaging the network security system completely. DDoS detection techniques are two types: anomaly based and signature based detection. Machine learning techniques are used in anomaly based detection while network behaviors are used in signature based detection. The infected host is prevented on software-defined network by using Deep Neural Network (DNN) and the damage of network by these infected hosts are also calculated in this method. Hierarchical Task Analysis (HTA) technique is used in calculating the proposed system validation. The comparative study gives performance of DDoS attack detection method in terms of accuracy, recall and precision.