Identification and Evaluation of Deep Learning-Based Defense Strategies to Counter DDoS Attacks in Cloud Computing Environments

Distributed Denial of Service (DDoS) attacks are considered serious threats in the field of cloud computing and have always been a concern for system administrators. In the present study, the main goal is to provide deep learning-based defensive strategies to mitigate the effects of DDoS attacks in cloud computing environments. Based on previous studies, a model with five inputs including packet size, flow rate, flow duration, TCP flags, and byte volume, was extracted. For the implementation of the proposed model, deep learning algorithms, specifically Long Short-Term Memory (LSTM) neural network, Recurrent Neural Network (RNN), and Deep Neural Network (DNN), were used. Evaluation results showed that the LSTM algorithm has the best performance with ۹۵% accuracy, followed by the RNN network with ۹۳% and the DNN network with ۹۲% accuracy. Also, among the input variables, flow rate had the most significant impact on attack detection, followed by packet size, byte volume, TCP flags, and flow duration in subsequent ranks of importance.

کلیدواژه ها:

Defense strategies ، Deep Learning ، DDoS attacks ، Cloud computing ، attack tools