The study of traffic anomalies using self-similar traffic model

The self-similarity becomes one of the method of anomaly detection. The multi-layer `on-off' model of traffic source is developed. The generator implemented according this model allows generate traffic flow characterized by self-similarity. The ability of generation of traffic similar to malicious one is verified by experiments. The results confirm that proposed model is useful to produce traffic similar to generated by the real attacker.