Publication Details
Detecting DoH-Based Data Exfiltration: FluBot Malware Case Study
DoH detection, malware detection, computer communication analysis, packet classification
This paper presents a novel approach for detecting the FluBot malware, an advanced Android banking Trojan that has been observed in active attacks in 2021 and 2022. The proposed method uses a two-layer detection mechanism to identify FluBot network connections. In the first layer, a machine learning algorithm is used to detect DNS-over-HTTPS (DoH) within Netflow records. The second layer uses a modified version of an existing domain generation algorithm (DGA) detection algorithm to target the DoH connections associated with the FluBot malware specifically. To evaluate the effectiveness of this approach, we used a dataset consisting of FluBot network traffic captured in a controlled sandbox environment. The preliminary results show that our DoH classifier achieves high accuracy and detection rates in identifying instances of FluBot malware, while maintaining a low false positive rate.
@inproceedings{BUT184570,
author="Roman {Rader} and Kamil {Jeřábek} and Ondřej {Ryšavý}",
title="Detecting DoH-Based Data Exfiltration: FluBot Malware Case Study",
booktitle="IEEE 48th Conference on Local Computer Networks (LCN)",
year="2023",
pages="50--54",
publisher="IEEE Computer Society",
address="Daytona Beach",
doi="10.1109/LCN58197.2023.10223341",
isbn="979-8-3503-0074-1",
url="https://www.fit.vut.cz/research/publication/13007/"
}