TRUSTCOM: Packet Trace Recognition over Unstable Communication Channels

Research Project, 2023 – 2024

Packet trace analysis uses information such as the packet sizes and their time of arrival to infer potentially private information from eavesdropped communication. The goal for this research project is to extend a passive network attack demonstrated in 2022 where the attacker was capable to recognize which video was being watched by the victim only based on analyzing the encrypted packet network traces. The attack was shown on the SVT Play platform by building a database containing information about all videos hosted by the web service. The following research questions are proposed to be further investigated:
• Can the attack be reproduced but using WPA2 encrypted link-layer traffic?
• How stable the packet trace needs to be for the attack to work?
• Can similar attacks work with other types of data?
The project intends to extend and test the already developped sniffer and recognition software with further experiments on WiFi connected devices such as smart TVs. A proposed approach for the extension is to develop a generic DASH video recognizer and filter out most of the noise in the packet traces using existing patterns in the network traffic. As a generalization of the attack, we propose to investigate further the recognition of packet data produced by other bust-like protocols.