Self‐similarity based DDoS attack detection using Hurst parameter

Abstract

AbstractDistributed denial‐of‐service attack is a serious concern in this era of voluminous Internet world. The challenge is to find difference between distributed denial‐of‐service attack traffic of all types and legitimate traffic at the earliest, as we can see high‐rate attack traffic and low rate attack traffic resembles with bursty legitimate traffic and normal legitimate traffic, respectively. Self‐similarity exists in ethernet traffic. Using Hurst parameter, we have differentiated legitimacy of any traffic irrespective of network traffic protocol. We have validated our method using both benchmark and real‐life datasets, and the results are highly satisfactory. Copyright © 2016 John Wiley & Sons, Ltd.