On a novel method of designing distributed LDoS attack based on network behavior features

Abstract

This paper designs a distributed low-rate denial of service (LDoS) attack using one-dimensional random walk algorithm. One-dimensional random walk algorithm can be used in generating power-law distribution which complies with normal network behavior feature. In this attack pattern, the behavior of each distributed attack flows are normal and nonperiodic, while the aggregated effect has the same damage degree as traditional square wave LDoS attacks. As the distributed attack flow satisfies the network traffic behavior feature, it is even more difficult to detect and trace back attack sources. The study on the feature of this attack provides reference for distributed low-rate DoS attack countermeasures.