Impact of network traffic anonymization on prediction and detection of DDoS attacks

Abstract

Effective prediction and detection of DDoS attacks directly depend on the quality of the information sources used as input for the prediction and detection mechanisms employed. Without a doubt, one of the best sources of information is the network traffic captured on the network interfaces of the protected devices. However, due to recent regulations such as the General Data Protection Regulation (LGPD) and the General Data Protection Regulation (GDPR), some traffic information needs to be anonymized, especially if the prediction and detection mechanisms are executed outside the domains. of the owner of the devices under protection, as in the case of computing cloud providers, or if the traffic is publicly shared with the scientific community with the aim of advancing the state of the art in the topic of cybersecurity. In this scientific initiation project, we intend to analyze how the anonymization of traffic affects the quality of the prediction and detection mechanisms for DDoS attacks created in the MENTORED project. (AU)