Chaos analysis based on Markov chains in DDoS detection

摘要

DDoS attacks would bring damage to the network. In traffic, Protocol is a useful packet field to represent the flow behavior. In this paper, a Markov chain is used to model the traffic pattern, based on protocol distribution. Then the global entropy rate and local entropy rate are calculated to detect the anomaly traffic in each time interval. Finally, chaotic analysis is applied based on the entropy rate time series. Experiment results show that the method can be effectively used to detect anomalies in network traffic.