DeepDeter: Strengthening Cybersecurity Against DoS Attacks with Deep Learning
Vanshika Pahuja, Sharad Shyam Ojha · 2024
Ever-evolving nature of cyber world has led to the significant increase in the different types of network attacks. Deep learning algorithms can be used for the detection of such types of attacks in which attackers have send many requests on the server and flooded it with the ample of requests and data. Many network packets have been sent to the targeted system which has led to the unavailability of online systems and crashing of sites. These attacks can cause the huge amount of loss to the systems and many online systems and targeted systems will become inoperative. This research work deploys deep learning techniques for the purpose of removal of DoS attacks and these deep learning techniques are Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) and Gated Recurrent Unit. These all techniques are used for the purpose of capturing of sequential data in the time-series data which will analyse all the data and identify patterns such as network patterns which will associate with the DoS attacks. LSTM outperforms best among all these but this model gives the highest accuracy of 92.3% which will indicate more superior ability for properly classification of all the instances of the attack traffic at the server.