Wireless Traffic Dataset for Krack and Kr00k Attacks in WPA2

This report centers on the “Wireless Traffic Dataset for KRACK and Kr00k Attacks in WPA2,” addressing the critical issues of data scarcity and poor reproducibility in this field. It systematically elaborates on the security mechanisms of WPA2 and the principles of its vulnerabilities. By constructing a realistic testbed with diverse clients, the study simulates and captures the complete traffic of both attacks. Moving beyond reliance on the direct release of raw traffic data, the report proposes a systematic framework for feature engineering and data preprocessing, resulting in the WPA2-KKID dataset comprising over 5.5 million data frames and 34 carefully selected features. The effectiveness of this dataset for attack detection was validated using machine learning methods, including SMOTE oversampling, PCA dimensionality reduction, and Random Forest. Finally, the report discusses the limitations of the current study regarding hardware/software scope and attack variety, outlining future directions for incorporating WPA3 and other vulnerabilities.