Fusion of Artificial Neural Networks by Fuzzy Logic Based Attack Detection Method

EasyChair Preprint 7023

Abstract

Increasing complexity and frequency of DNS attacks have increased the need for attacks detection and incident response. So new ways are needed to perform traffic analysis for Intrusion Detection Systems (IDS) in DNS network. In this paper, a Deniel of Service (DOS) attack security situation assessment model using fusion feature based on Fuzzy C-means (FCM) clustering algorithm has been proposed. This model generates a fusion feature according to network flow changes in IP address domain name of each connection DNS, and calculates the risk index of each DNS network node on the basis of fusion feature and obtains the security situation information of the whole network by fusing the risk indexes of all network nodes, and clusters the fusion situation information with FCM into five security levels, so as to quantitatively evaluate the DoS attack security situation of the whole DNS network through the proposed situation risk degree recognition model. Experiments on real DoS data show that the proposed model can assess the DoS attack security situation reasonably and effectively and be more flexible than non-fusion based methods

Keyphrases: DDoS attack, DNS protocol, FCM, fusion feature

@booklet{EasyChair:7023,
  author    = {Adel Dallali and Takwa Omrani and Belgacem Chibani Rhaimi},
  title     = {Fusion of Artificial Neural Networks by Fuzzy Logic Based Attack Detection Method},
  howpublished = {EasyChair Preprint 7023},
  year      = {EasyChair, 2021}}