Collaborative detection of DDoS attacks based on chord protocol

This paper propose a new method for detecting distributed denial of service (DDoS) attack. By using the monitors on the edge network to record the traffic information in a Sketch matrix, we can get the suspicious attack traffic information as the local detection result. Then, all the monitors that have found suspicious information about the same destination IP organize into an AutoTree to collaborative aggregate the local detection result. The local detection information will be aggregated in the AutoTree step by step, while the root of the AutoTree will be in charge of the final decision for the target confirmation. In order to verify the proposed method, we do the experiments in a simulated network. We compare the detection time, detection rate between three detection methods: Collaborative Distributed Detection (CDD), Routing Distribution Based Detection (RDBD) and Single Point Detection (SPD).