An approach to detecting anomalies in a self-similar network traffic

Aim. The effects of cyber attacks cause failures network elements, theft information and other unlawful actions. Cyber are often accompanied by untypical traffic activity anomalies. paper aims to develop an approach detecting anomalies in identifying the degree self-similarity using fractal analysis statistical methods. Methods. uses methods mathematical statistics, analysis, analysis. Results. suggests evaluating for improving accuracy attack detection. At first stage, Hurst exponent is calculated reference traffic. second actual divided into optimal time intervals; each interval, calculated. If identified value differs from one obtained traffic, it decided that there anomaly. final used order precisely localise authors analysed resulted identification more efficient be as part proposed approach. For DFA method was proposed, while ARFIMA proposed. Conclusion. suggested allows real or near-real time.