TOR NETWORK VIDEO FINGERPRINTING OVER RESIDENTIAL WI-FI AND CONGESTED NETWORK INTERFACES

Both ordinary, concerned citizens and malicious actors, such as extremist groups, use the Onion Router (Tor) to encrypt their web traffic and protect their online anonymity. This usage has led to an interest in attacks on Tor, such as video fingerprinting, which seeks to identify what video is being streamed solely by analyzing features in the encrypted traffic trace. This thesis seeks to continue a previous line of work in "Exploring the Capabilities and Limitations of Video Stream Fingerprinting" to use Convolutional Neural Networks to conduct video fingerprinting attacks on network traffic transmitted over Tor in an open-world environment. It will expand upon previous research using a realistic vantage point to collect data utilizing Wi-Fi and traffic over the network interface. The end goal of this line of research is to increase the Department of Defense's understanding of how much risk exists to the privacy of the users of the Tor network. While performance suffers compared to models trained on data sets collected under ideal conditions, we show that video fingerprinting is still viable in typical network conditions, including robustness to traffic congestion in all but the most extreme cases.

Department dc:contributor.department

Computer Science (CS)

Grantor dc:publisher

Monterey, CA; Naval Postgraduate School

Year dc:date.issued

2024

Author dc:creator

• Falk, Theodore B.

Advisors dc:contributor.advisor

• Barton, Armon C.
• Walsh, Timothy C.

Statement dc:rights

• This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.

Handle dc:identifier.uri

https://hdl.handle.net/10945/73449

OAI identifier oai:identifier

oai:calhoun.nps.edu:10945/73449