liu.seSearch for publications in DiVA
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Understanding and Improving Video Fingerprinting Attack Accuracy under Challenging Conditions
Linköping University, Department of Computer and Information Science, Database and information techniques. Linköping University, Faculty of Science & Engineering.
Linköping University, Department of Computer and Information Science, Database and information techniques. Linköping University, Faculty of Science & Engineering. Sectra Communications, Linköping, Sweden.
Linköping University, Department of Computer and Information Science, Database and information techniques. Linköping University, Faculty of Science & Engineering.ORCID iD: 0009-0006-8038-5693
Linköping University, Department of Electrical Engineering, Information Coding. Linköping University, Faculty of Science & Engineering. Sectra Communications, Linköping, Sweden.ORCID iD: 0000-0001-5888-1291
Show others and affiliations
2024 (English)In: PROCEEDINGS OF THE 23RD WORKSHOP ON PRIVACY IN THE ELECTRONIC SOCIETY, WPES 2024, ASSOC COMPUTING MACHINERY , 2024, p. 141-154Conference paper, Published paper (Refereed)
Abstract [en]

The threat of video fingerprinting attacks poses significant privacy concerns. These attacks can identify streamed videos with high accuracy despite the use of encryption, leveraging both heuristic-based and deep learning techniques. However, the real-world effectiveness of such attacks remains underexplored, as most research assumes ideal conditions. In this paper, we address the challenges posed by variable network conditions and live-streaming latency, which complicate the attacker's ability to collect useful training data. First, we evaluate several deep learning model architectures against video data under diverse network conditions, including two adaptations of existing website fingerprinting attacks tailored to video that we show boast notable improvements over the base attacks and previous state-of-the-art video fingerprinting attacks. Second, we introduce two augmentation techniques and demonstrate that they substantially enhance attack performance in suboptimal conditions, without knowledge of the victim's live latency. Finally, we analyze the effects of data limitations such as observation time, dataset size, and training time. Overall, our work provides new insights into the impact that several real-world challenges have on attack accuracy, presents new and improved attacks, and details two augmentation techniques that can further boost the performance of the new attacks. Combined, these significant advancements highlight the urgent need for effective defense mechanisms.

Place, publisher, year, edition, pages
ASSOC COMPUTING MACHINERY , 2024. p. 141-154
Keywords [en]
traffic analysis; video fingerprinting; challenging conditions
National Category
Other Engineering and Technologies
Identifiers
URN: urn:nbn:se:liu:diva-212753DOI: 10.1145/3689943.3695045ISI: 001434853500011Scopus ID: 2-s2.0-85214235108ISBN: 9798400712395 (print)OAI: oai:DiVA.org:liu-212753DiVA, id: diva2:1949233
Conference
23rd Workshop on Privacy in the Electronic Society, Salt Lake City, UT, oct 14-18, 2024
Note

Funding Agencies|Swedish Foundation for Strategic Research (SSF); Wallenberg AI, Autonomous Systems and Software Program (WASP) - Knut and Alice Wallenberg Foundation

Available from: 2025-04-02 Created: 2025-04-02 Last updated: 2025-09-01
In thesis
1. Toward Secure and Privacy-Preserving Communication over Non-Trusted Networks
Open this publication in new window or tab >>Toward Secure and Privacy-Preserving Communication over Non-Trusted Networks
2025 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

The widespread adoption of encrypted communication protocols has improved digital privacy. However, even if the content is encrypted, an attacker can exploit metadata and patterns in network traffic to infer a user’s activity and behavior. Meanwhile, as users increasingly rely on third-party infrastructure, cloud-based applications face privacy challenges during data processing, exposing sensitive information. In parallel, long-term communication security depends on proper certificate management, where misconfigurations or evolving practices can compromise security. This thesis explores these multifaceted challenges and presents solutions to preserve privacy in adversarial and non-trusted environments.

First, the thesis focuses on encrypted traffic analysis, particularly fingerprinting at-tacks that exploit observable metadata such as packet sizes, transmission timing, and traffic flow patterns to infer sensitive user information and identify user activities. The work explores these attacks across multiple platforms and use cases, demonstrating their real-world feasibility and high accuracy. To counteract these threats, several mitigation strategies are systematically evaluated, including packet padding, timing obfuscation, and traffic shaping, each assessed for its tradeoffs between effectiveness, impact on network performance, and users’ quality of experience.

Second, the thesis studies secure data computation in cloud environments using homomorphic encryption (HE), a cryptographic technique that enables computation directly on encrypted data without requiring prior decryption. While HE offers a strong theoretical foundation, its practical application has long been hindered by performance overhead and integration complexity. This thesis explores the real-world applicability of HE by designing diverse systems, comparing schemes, and proposing efficiency optimizations. The findings highlight the potential and current limitations of HE, offering valuable guidance for its adoption in cloud-based systems.

Third, the thesis examines long-term authentication security through a 10-year longitudinal analysis of certificate usage in the web public key infrastructure. While certificates are essential for encrypted communication, inconsistent issuance, renewal, and management can introduce systemic vulnerabilities. The analysis of wildcard and multi-domain certificates, as well as certificate chain evolution, reveals key trends, including declining use of wildcard certificates, shifting practices among certificate authorities, and simpler chain structures. These patterns highlight evolving industry behaviors and persistent challenges in certificate lifecycle management.

Combined, this thesis contributes to a better understanding of the evolving security and privacy landscape in digitally connected systems. By contributing to three distinct but interrelated domains, the thesis highlights the complexity of modern privacy challenges and offers targeted strategies to strengthen digital confidentiality. Through systematic evaluations, novel designs, and long-term measurements, the work advances state-of-the-art privacy-preserving communication and provides practical insights for building a more resilient and trustworthy digital infrastructure.

Abstract [sv]

I dagens digitala värld används kryptering i allt större utsträckning för att skydda vår kommunikation och stärka den digitala integriteten. Men även om innehållet är krypterat kan en angripare utnyttja metadata och mönster i nätverkstrafik för att dra slutsatser om en användares aktivitet och beteenden. Samtidigt har användningen av molntjänster ökat kraftigt, vilket ställer höga krav på att databehandlingen sker på säkert sätt då känslig information ofta behandlas på tredjepartsinfrastruktur. Dessutom är långsiktig säkerhet i digital kommunikation beroende av korrekt hantering av digitala certifikat, där felkonfigurationer eller föråldrade metoder kan skapa säkerhetsluckor. Den här avhandlingen undersöker dessa komplexa och sam-manlänkade utmaningar samt presenterar lösningar för att skydda integriteten i miljöer där nätverk eller tjänsteleverantörer inte kan betraktas som tillförlitliga.

Först studeras hur angripare kan utföra trafikanalysattacker på krypterad nätverks-trafik genom att analysera metadata, till exempel paketstorlekar, tidsintervaller och trafikflöden. Med mönsteranalys blir det möjligt att med hög noggrannhet fastställa en användares aktivitet, även när innehållet är krypterat. Avhandlingen utforskar dessa attacker i olika områden och demonstrerar dess praktiska genomförbarhet. För att motverka attackerna utvärderas även olika skyddsåtgärder, som att skicka med extra data, fördröja sändningar eller forma om trafiken, alla som medför olika avvägningar mellan skyddsnivå, nätverksprestanda och användarupplevelse.

Avhandlingen studerar även säker databehandling i molnet med hjälp av homomorfisk kryptering, en kryptografisk teknik som möjliggör beräkningar direkt på krypterad data utan att den först behöver dekrypteras. Detta innebär att molntjänster kan behandla information utan att behöva ha tillgång till den i klartext. Trots att tekniken är lovande, har dess praktiska tillämpning försvårats av stora utmaningar i form av höga prestandakrav och komplex integrering. Avhandlingen undersöker teknikens tillämpbarhet genom olika systemdesign och algoritmjämförelser, vilket belyser både potential och begränsningar samt ger insikter för praktisk användning.

Slutligen analyseras hur det digitala certifikatsystemet, som utgör grunden för säker autentisering på webben, har utvecklats under de senaste tio åren. Genom omfattande dataanalys identifieras förändrade mönster i hur certifikat utfärdas, förnyas och hanteras. Resultaten visar på både positiva förändringar och kvarstående utmaningar, samt betonar behovet av tydligare riktlinjer och förbättrade rutiner för att långsiktigt stärka tilliten till den digitala infrastrukturen.

Sammanfattningsvis bidrar avhandlingen till en fördjupad förståelse av säkerhet och integritet i digitala kommunikationssystem. Genom systematiska analyser, praktiska experiment och långsiktiga mätningar presenteras lösningar som gör dagens och morgondagens digitala infrastruktur mer robust och tillförlitlig.

Place, publisher, year, edition, pages
Linköping: Linköping University Electronic Press, 2025. p. 103
Series
Linköping Studies in Science and Technology. Dissertations, ISSN 0345-7524 ; 2461
National Category
Security, Privacy and Cryptography
Identifiers
urn:nbn:se:liu:diva-217111 (URN)10.3384/9789181181838 (DOI)9789181181821 (ISBN)9789181181838 (ISBN)
Public defence
2025-10-03, Ada lovelace, B-building, Campus Valla, Linköping, 09:15 (English)
Opponent
Supervisors
Note

Funding: This work was partially supported by the Wallenberg AI, Autonomous Systems and Software Program (WASP) funded by the Knut and Alice Wallenberg Foundation.

2025-09-18: Updated to a smaller file size.

Available from: 2025-09-01 Created: 2025-09-01 Last updated: 2025-09-18Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textScopus

Search in DiVA

By author/editor
Carlson, AugustHasselquist, DavidWitwer, EthanJohansson, NiklasCarlsson, Niklas
By organisation
Database and information techniquesFaculty of Science & EngineeringInformation Coding
Other Engineering and Technologies

Search outside of DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 103 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • oxford
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf