Network traffic analysis - a different approach using incoming and outgoing traffic differences

Abstract

The network traffic analysis is a fundamental area on network management because the network anomalies may affect the network quality of service. However, the data network traffic anomalies are still a critical issue. On last years signal processing methods like wavelet-based ones have been used to detect anomalies on network traffic, specially because wavelet transforms allow the selection of signal characteristics via a combined time-frequency representation. This paper explores a simple and fast wavelet transform for analyzing the network flow, considering the difference between incoming and outgoing traffic data, for improving identification of deny of service attacks.