A New Method of Testing Machine Learning Models of Detection for Targeted DDoS Attacks

Mateusz Kozlowski, Bogdan Ksiezopolski

2021

Abstract

Distributed Denial of Service (DDoS) is one of the most popular attacks on the Internet. One of the most popular classes of DDoS attacks is the flood-based, which sends huge amounts of packets to the victim host or infrastructure, causing an overload of the system. One of the attack mitigation systems is based on machine learning (ML) methods, which in many cases has a very high accuracy rate (0.95 – 0.99). Unfortunately, most ML models are not resistant against targeted DDoS attacks. In this article, we present the targeted attacks to the DDoS ML-based mitigation models, which have a high accuracy. After this, we propose a new method of testing ML-based models against targeted DDoS attacks.

Paper Citation

in Harvard Style

Kozlowski M. and Ksiezopolski B. (2021). A New Method of Testing Machine Learning Models of Detection for Targeted DDoS Attacks. In Proceedings of the 18th International Conference on Security and Cryptography - Volume 1: SECRYPT, ISBN 978-989-758-524-1, pages 728-733. DOI: 10.5220/0010574507280733

in Bibtex Style

@conference{secrypt21,
author={Mateusz Kozlowski and Bogdan Ksiezopolski},
title={A New Method of Testing Machine Learning Models of Detection for Targeted DDoS Attacks},
booktitle={Proceedings of the 18th International Conference on Security and Cryptography - Volume 1: SECRYPT,},
year={2021},
pages={728-733},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010574507280733},
isbn={978-989-758-524-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 18th International Conference on Security and Cryptography - Volume 1: SECRYPT,
TI - A New Method of Testing Machine Learning Models of Detection for Targeted DDoS Attacks
SN - 978-989-758-524-1
AU - Kozlowski M.
AU - Ksiezopolski B.
PY - 2021
SP - 728
EP - 733
DO - 10.5220/0010574507280733