Enhancing Network Security: A Machine Learning-Based Approach for Detecting and Mitigating Krack and Kr00k Attacks in IEEE 802.11

Zaher Salah, Esraa Abu Elsoud

The rise in internet users has brought with it the impending threat of cybercrime as the Internet of Things (IoT) increases and the introduction of 5G technologies continues to transform our digital world. It is now essential to protect communication networks from illegal intrusions to guarantee data integrity and user privacy. In this situation, machine learning techniques used in data mining have proven to be effective tools for constructing intrusion detection systems (IDS) and improving their precision. We use the well-known AWID3 dataset, a comprehensive collection of wireless network traffic, to investigate the effectiveness of machine learning in enhancing network security. Our work primarily concentrates on Krack and Kr00k attacks, which target the most recent and dangerous flaws in IEEE 802.11 protocols. Through diligent implementation, we were able to successfully identify these threats using an IDS model that is based on machine learning. Notably, the resilience of our method was demonstrated by our ensemble classifier’s astounding 99% success rate in detecting the Krack attack. The effectiveness of our suggested remedy was further demonstrated by the high accuracy rate of 96.7% displayed by our neural network-based model in recognizing instances of the Kr00k attack. Our research shows the potential for considerably boosting network security in the face of new threats by leveraging the capabilities of machine learning and a diversified dataset. Our findings open the door for stronger, more proactive security measures to protect IEEE. 802.11 networks’ integrity, resulting in a safer online environment for all users.