A DDoS Attack Detection Method Using Conditional Entropy Based on SDN Traffic

Qiwen Tian, Sumiko Miyata

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

To detect each network attack in an SDN environment, an attack detection method is proposed based on an analysis of the features of the attack and the change in entropy of each parameter. Entropy is a parameter used in information theory to express a certain degree of order. However, with the increasing complexity of networks and the diversity of attack types, existing studies use a single entropy, which does not discriminate correctly between attacks and normal traffic and may lead to false positives. In this paper, we propose new state determination standards that use the normal distribution characteristics of the entropy value at the time which an attack did not occur, subdivide the normal and abnormal range represented by the entropy value, improving the accuracy of attack determination. Furthermore, we show the effectiveness of the proposed method by numerical analysis.

Original languageEnglish
Pages (from-to)95-111
Number of pages17
JournalInternet of Things
Volume4
Issue number2
DOIs
Publication statusPublished - 2023 Jun

Keywords

  • abnormal traffic
  • attack detection
  • DDoS
  • entropy
  • flash crowds
  • SDN

ASJC Scopus subject areas

  • Computer Science (miscellaneous)
  • Engineering (miscellaneous)
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'A DDoS Attack Detection Method Using Conditional Entropy Based on SDN Traffic'. Together they form a unique fingerprint.

Cite this