Advanced Cyber Attack Detection and Mitigation in Software-Defined Networking: A Deep Kerberos Authentication-Based Gradient Chaos Aquila Optimization Approach
M. Priyadharshini, B. Vanathi · Cybernetics & Systems · 2026
Software-defined networking provides quickness and adaptability in network management, which allows improved efficiency and minimized functional costs. Nevertheless, the rising incidence of cyber attacks, especially distributed denial of service attacks, shows a considerable attacks on network security. On the other hand, deep learning approaches incorporated with software-defined networking have demonstrated promising results in mitigating these attacks. Meanwhile, conventional approaches face drawbacks such as reliance on static rules, low detection rates, computational inefficiency, and delayed notifications. To rectify these difficulties, this paper proposes the Deep Kerberos Authentication-based Gradient Chaos Aquila Optimization algorithm. The network attack data is collected from three datasets, which are the datasets used in this work. In this work, handling missing values, feature scaling, one-hot encoding, and balanced random sampling are the diverse preprocessing techniques used to preprocess the data. The Gradient Chaos Aquila Optimization algorithm is used in the feature selection process, which is the combination of Aquila Optimization, Gradient Search Rule, and Chaos Optimization. This algorithm rectifies local optima issues and balances exploration and exploitation, thus improving detection accuracy. The experimental results showed better performance in terms of classification accuracy with 98.62% and throughput with 2650 R/S from proposed model rather than existing models.