Walls Have Ears: Traffic-based Side-channel Attack in Video Streaming

Gu, Jiaxi; Wang, Jiliang; Yu, Zhiwen; Shen, Kele

doi:10.1109/infocom.2018.8486211

Cited by 38 publications

(24 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In particular [18] focused on identifying video streaming using a Variable Bit Rate algorithm within encrypted WiFi traffic using similarity metrics and statistical machine learning. In a similar manner, but with the addition of DASH streaming, [19] identified video streaming within WiFi traffic. In their latest work [19], the authors adopted an approach similar to [18] by modelling various video streaming traffic.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Deep Content: Unveiling Video Streaming Content from Encrypted WiFi Traffic

Liu

Huang

et al. 2018

2018 IEEE 17th International Symposium on Network Computing and Applications (NCA)

View full text Add to dashboard Cite

The proliferation of smart devices has led to an exponential growth in digital media consumption, especially mobile video for content marketing. The vast majority of the associated Internet traffic is now end-to-end encrypted, and while encryption provides better user privacy and security, it has made network surveillance an impossible task. The result is an unchecked environment for exploiters and attackers to distribute content such as fake, radical and propaganda videos.Recent advances in machine learning techniques have shown great promise in characterising encrypted traffic captured at the end points. However, video fingerprinting from passively listening to encrypted traffic, especially wireless traffic, has been reported as a challenging task due to the difficulty in distinguishing retransmissions and multiple flows on the same link. We show the potential of fingerprinting videos by passively sniffing WiFi frames in air, even without connecting to the WiFi network. We have developed Multi-Layer Perceptron (MLP) and Recurrent Neural Networks (RNNs) that are able to identify streamed YouTube videos from a closed set, by sniffing WiFi traffic encrypted at both Media Access Control (MAC) and Network layers. We compare these models to the state-ofthe-art wired traffic classifier based on Convolutional Neural Networks (CNNs), and show that our models obtain similar results while requiring significantly less computational power and time (threefold reduction).

show abstract

Section: Related Workmentioning

confidence: 99%

“…Recent research has successfully identified video streaming content in encrypted traffic without using deep learning techniques [18], [19]. In particular [18] focused on identifying video streaming using a Variable Bit Rate algorithm within encrypted WiFi traffic using similarity metrics and statistical machine learning.…”

Section: Related Workmentioning

confidence: 99%

Deep Content: Unveiling Video Streaming Content from Encrypted WiFi Traffic

Liu

Huang

et al. 2018

2018 IEEE 17th International Symposium on Network Computing and Applications (NCA)

View full text Add to dashboard Cite

show abstract

“…A number of studies [9]- [13] have been conducted to identify video titles watched by clients and they use mainly machine learning methods to distinguish video traffic features that change over time. In another study, a technique called dynamic time warping was used to measure the similarity between video fingerprints or signatures calculated using video packet length [17]. However, these studies are inadequate in real-world applications because they assume that the network condition is stable such that the same traffic pattern is always reproduced for the streaming of one video.…”

Section: Introductionmentioning

confidence: 99%

VTIM: Video Title Identification Using Open Metadata

et al. 2020

View full text Add to dashboard Cite

Video-streaming applications are very popular these days. Existing studies of video streaming have attempted to identify video titles of users using machine learning techniques to identify specific patterns of video packets transmitted over the network. However, these studies have limitations when applied to actual environments where the network is congested or there are multiple users in the same network. This paper proposes Video Title Identification using open Metadata (VTIM), which identifies video titles by analyzing storyboards and Media Presentation Description (MPD) of MPEG-DASH in connection with video packets transmitted over the network. Attack was carried out using VTIM on 13,291 videos selected from actual video-streaming environment of YouTube. Our experiments show that VTIM is able to identify video titles with 100% accuracy at nearly thirty times faster than existing methods based on machine learning techniques. The paper also proposes and evaluates a countermeasure against VTIM.

show abstract

“…Several prior works [5,7,12,15] have examined the fingerprintability of video streams under typical browsing conditions (HTTPS).…”

Section: Introductionmentioning

confidence: 99%

Poster

Rahman

Matthews

Wright

2019

Proceedings of the 2019 ACM SIGSAC Conference on Computer and Communications Security

View full text Add to dashboard Cite

Over 8 million users rely on the Tor network each day to protect their anonymity online. Unfortunately, Tor has been shown to be vulnerable to the website fingerprinting attack, which allows an attacker to deduce the website a user is visiting based on patterns in their traffic. The state-of-the-art attacks leverage deep learning to achieve high classification accuracy using raw packet information. Work thus far, however, has examined only one type of media delivered over the Tor network: web pages, and mostly just home pages of sites. In this work, we instead investigate the fingerprintability of video content served over Tor. We collected a large new dataset of network traces for 50 YouTube videos of similar length. Our preliminary experiments utilizing a convolutional neural network model proposed in prior works has yielded promising classification results, achieving up to 55% accuracy. This shows the potential to unmask the individual videos that users are viewing over Tor, creating further privacy challenges to consider when defending against website fingerprinting attacks.

show abstract

Walls Have Ears: Traffic-based Side-channel Attack in Video Streaming

Cited by 38 publications

References 24 publications

Deep Content: Unveiling Video Streaming Content from Encrypted WiFi Traffic

Deep Content: Unveiling Video Streaming Content from Encrypted WiFi Traffic

VTIM: Video Title Identification Using Open Metadata

Poster

Contact Info

Product

Resources

About