DDoS Attack Detection Based on Differential Privacy Graph Data Protection in IoV

With IoV's rapid growth and widespread use, securing it has become a key focus in scholarly research.In vehicle networking, DDoS attacks are prevalent and impactful.Existing deep learning methods for DDoS detection in IoV often yield superior results.However, these methods typically analyze individual flows, neglecting flow relationships and resulting in poor robustness.In addition,given the sensitivity of communication data containing substantial information, ensuring its security is imperative.This study proposes a DDoS detection method for IoV based on differential privacy graph data protection.The method detects whether a DDoS attack occurs by capturing vehicle networking communication flows and analyzing their relationships.We also employ the differential privacy technique based on attribute importance and feature relevance to protect IoV graph data at varying privacy levels.It is proved through experiments that the method can have the ability to get high accuracy of attack detection while ensuring privacy.