|
A Review on Website Fingerprinting Attacks
|
Maryam Taebi   , Ali Bohlooli * , Marjan Kaedi |
| University of Isfahan |
|
|
Abstract: (3004 Views) |
In Website Fingerprinting (WFP) Attacks, clients’ destination webpages are identified using traffic analysis techniques, without any need to decrypt traffic contents. Typically, clients make use of the privacy enhancing technologies (e.g., VPNs, proxies, and anonymity networks) to browse webpages. These technologies allow clients to hide traffic contents and their real destinations. To perform an attack, features are extracted from the input packet sequence. Next, the data is pre-processed and finally, client’s real destination is revealed by means of a machine learning algorithm. Various studies have utilized statistical methods or classification approaches to infer the client’s visited webpages. In this paper, a comprehensive overview of WFP techniques is performed, in which previous studies are categorized based on the features they use for webpages identification. This is a new approach for categorizing previous works on WFP attacks and to the best of our knowledge, this viewpoint has not been applied so far.
|
|
| Keywords: Website Fingerprinting Attack, Traffic Analysis, Machine Learning, Classification, Feature Extraction |
|
|
Full-Text [PDF 3981 kb]
(798 Downloads)
|
Type of Study: Review Article |
Subject:
Special
Received: 2017/11/16 | Accepted: 2018/06/10 | Published: 2018/08/27
|
|
|
|
|
|
