Computer Laboratory

Technical reports

Characterization of Internet censorship from multiple perspectives

Sheharbano Khattak

January 2017, 170 pages

This technical report is based on a dissertation submitted January 2017 by the author for the degree of Doctor of Philosophy to the University of Cambridge, Robinson College.


Internet censorship is rampant, both under the support of nation states and private actors, with important socio-economic and policy implications. Yet many issues around Internet censorship remain poorly understood because of the lack of adequate approaches to measure the phenomenon at scale. This thesis aims to help fill this gap by developing three methodologies to derive censorship ground truths, that are then applied to real-world datasets to study the effects of Internet censorship. These measurements are given foundation in a comprehensive taxonomy that captures the mechanics, scope, and dynamics of Internet censorship, complemented by a framework that is employed to systematize over 70 censorship resistance systems.

The first part of this dissertation analyzes “user-side censorship”, where a device near the user, such as the local ISP or the national backbone, blocks the user’s online communication. This study provides quantified insights into how censorship affects users, content providers, and Internet Service Providers (ISPs); as seen through the lens of traffic datasets captured at an ISP in Pakistan over a period of three years, beginning in 2011.

The second part of this dissertation moves to “publisher-side censorship”. This is a new kind of blocking where the user’s request arrives at the Web publisher, but the publisher (or something working on its behalf) refuses to respond based on some property of the user. Publisher-side censorship is explored in two contexts. The first is in the context of an anonymity network, Tor, involving a systematic enumeration and characterization of websites that treat Tor users differently from other users.

Continuing on the topic of publisher-side blocking, the second case study examines the Web’s differential treatment of users of adblocking software. The rising popularity of adblockers in recent years poses a serious threat to the online advertising industry, prompting publishers to actively detect users of adblockers and subsequently block them or otherwise coerce them to disable the adblocker. This study presents a first characterization of such practices across the Alexa Top 5,000 websites.

This dissertation demonstrates how the censor’s blocking choices can leave behind a detectable pattern in network communications, that can be leveraged to establish exact mechanisms of censorship. This knowledge facilitates the characterization of censorship from different perspectives; uncovering entities involved in censorship, its targets, and the effects of such practices on stakeholders. More broadly, this study complements efforts to illuminate the nature, scale, and effects of opaque filtering practices; equipping policy-makers with the knowledge necessary to systematically and effectively respond to Internet censorship.

Full text

PDF (3.6 MB)

BibTeX record

  author =	 {Khattak, Sheharbano},
  title = 	 {{Characterization of Internet censorship from multiple
  year = 	 2017,
  month = 	 jan,
  url = 	 {},
  institution =  {University of Cambridge, Computer Laboratory},
  number = 	 {UCAM-CL-TR-897}