This file was created by the TYPO3 extension bib --- Timezone: CEST Creation date: 2024-05-06 Creation time: 01-43-13 --- Number of references 1 inproceedings 2021_mitseva_sequences 2021 11 17 2411-2413 Website fingerprinting (WFP) is a special case of traffic analysis, where a passive attacker infers information about the content of encrypted and anonymized connections by observing patterns of data flows. Although modern WFP attacks pose a serious threat to online privacy of users, including Tor users, they usually aim to detect single pages only. By ignoring the browsing behavior of users, the attacker excludes valuable information: users visit multiple pages of a single website consecutively, e.g., by following links. In this paper, we propose two novel methods that can take advantage of the consecutive visits of multiple pages to detect websites. We show that two up to three clicks within a site allow attackers to boost the accuracy by more than 20% and to dramatically increase the threat to users' privacy. We argue that WFP defenses have to consider this new dimension of the attack surface. Traffic Analysis; Website Fingerprinting; Web Privacy https://www.comsys.rwth-aachen.de/fileadmin/papers/2021/2021-mitseva-fingerprinting-sequences.pdf ACM Proceedings of the 28th ACM SIGSAC Conference on Computer and Communications Security (CCS '21), November 15-19, 2021, Seoul, Korea Seoul, Korea November 15-19, 2021 978-1-4503-8454-4/21/11 10.1145/3460120.3485347 1 AsyaMitseva JanPennekamp JohannesLohmöller TorstenZiemann CarlHoerchner KlausWehrle AndriyPanchenko