A real time detection of distributed denial-of-service attacks using cumulative sum algorithm and adaptive neuro-fuzzy inference system

Abstract

Distributed denial-of-service (DDoS) is a very powerful attack on Internet resources as well as system resources. Hence, it is imperative to detect these attacks in real time else the impact will be irresistible.In this work we propose a new method of applying cumulative sum (CUSUM) algorithm to track variations of the attack characteristic variable X(n) from the observed traffic (specific to different kinds of attacks) and raise an alarm based on threshold. But often a threshold based mechanism produces many false alarms. Adaptive Neuro Fuzzy Inference System (ANFIS) which is capable of removing the abrupt separation between normality and abnormality as well as appropriately select the membership function parameters has been used for detection of attacks based on CUSUM values. The detection mechanism is well corroborated by experimental results. © 2012 Springer-Verlag GmbH.