|
Abstract
|
Due to the increasing complexity of computer networks and increased attacks, the security of these networks has also been complicated and difficult. Hence, penetration prediction systems are provided as an active way to secure computer networks. The task of predictive penetration systems is the predictive analysis of the past behavior of the system. In this paper, we have tried to provide a method that can predict the future of the network and detect chaos theory by predicting DDoS attacks. With this prediction, we can detect early attacks on attacks. In this way, the future traffic of the network is predicted based on the model of exponential smoothing and based on the predicted error, the prediction error series is obtained from the actual traffic value. As shown, this series has chaotic features. It has also been shown that by using the Lyapunov component analysis, the probability of attacks can be detected in this time series. For this purpose, the resulting time series are predicted using the ESN method and using the Lyapunov component analysis, the attack is predicted.
To demonstrate the proposed method, we show with the Darpa98 dataset, which consists of collected data and a standard for evaluating intrusion detection systems, that this method has the ability to predict early attacks.
|