Distributed Denial of Service Attack Classification Using Artificial Neural Networks.

EasyChair Preprint 3201

Abstract

For all sizes of organizations and ISPs, most devastating attacks of all time are emerged by the DDoS Attacks (Distributed Denial of Service). The contribution of production of unsecured botnets and IoT devices in the range of billions number is lead to the increment of DDoS attacks due to the improved availability of services of DDoS-for-hire. Continuously, these DDoS attacks are growing in frequency, magnitude, and sophistication. Owing to the smarter growing of these attacks day by day and evasion of IDS, the legacy methods are challenged that include scrubbing and signature-based detection. As the scale of attacks mostly concentrating on the organizations, the security technologies of next-generation can’t keep in the pace. Due to the higher demand of human intervention, various limitations are included the anomaly-based detection with false positives and accuracy. By using machine learning (ML) model, DDoS anomaly detection based on the dataset of open CICIDS2017 is presented in this paper. However, maximum accuracy is reached with the use of this ML model and tuning hyper parameters meticulously.

Keyphrases: Accuracy, DDoS attacks, Intrusion Detection System, anomaly detection, machine learning

@booklet{EasyChair:3201,
  author    = {Bhargavi Goparaju and Srinivasa Rao Bandla},
  title     = {Distributed Denial of Service Attack Classification Using Artificial Neural Networks.},
  howpublished = {EasyChair Preprint 3201},
  year      = {EasyChair, 2020}}