A hierarchical approach to traffic anomaly detection using image processing technique

Increasing malicious network traffic has been serious threats to the network security and network administrators have difficulty to detect the network attacks from vast network traffic. Because an image can contain the much traffic information and intuitively display the network status, it is helpful to reduce the processing time for detecting the anomalies. Therefore we proposed a hierarchical approach to detecting various network attacks using a two-tiered system of image analysis. In a first tier, random attacks are detected by analyzing the global traffic and we will be able to discover semi-random attacks by examining the local traffic images in second tier. The proposed method can effectively detects small-scale attacks like scanning attacks as well as large-scale attacks such as DDos, Worm and etc.