Real-Time DDoS Detection and Alleviation in Software-Defined In-Vehicle Networks

In-vehicle network (IVN) is deployed in the autonomous car to assist the data transmission among sensors, electronic control units, and a server in taking care of the data processing and driving management. However, sensors might be compromised by attackers to flood a large amount of packets in the IVN. Under this circumstance, packet loss or large transmission delay might occur, which, in turn, reduces the driving safety. We propose a Joint K-means clustering and Software-defined networking removing framework to instantly detect and remove the suspicious sensors caused distributed denial of service (DDoS) attack in the IVN with the assistance of software-defined networking and machine learning. The proposed JKS is integrated into the existing network system and shows the potential to time efficiently detect and mitigate the DDoS attack in the IVN.