The Anatomy of Web Censorship in Pakistan

TL;DR

This study investigates the causes, mechanisms, and evolution of web censorship in Pakistan, revealing shifts in censorship infrastructure and exploring user circumvention techniques.

Contribution

First comprehensive analysis of Pakistan's web censorship mechanisms, including a comparison of two generations of blocking systems and user circumvention methods.

Findings

Censorship varies at DNS and HTTP levels.

Shift to centralized exchange-level censorship observed.

Users employ various methods to bypass restrictions.

Abstract

Over the years, the Internet has democratized the flow of information. Unfortunately, in parallel, authoritarian regimes and other entities (such as ISPs) for their vested interests have curtailed this flow by partially or fully censoring the web. The policy, mechanism, and extent of this censorship varies from country to country. We present the first study of the cause, effect, and mechanism of web censorship in Pakistan. Specifically, we use a publicly available list of blocked websites and check their accessibility from multiple networks within the country. Our results indicate that the censorship mechanism varies across websites: some are blocked at the DNS level while others at the HTTP level. Interestingly, the government shifted to a centralized, Internet exchange level censorship system during the course of our study, enabling our findings to compare two generations of blocking systems. Furthermore, we report the outcome of a controlled survey to ascertain the mechanisms that are being actively employed by people to circumvent censorship. Finally, we discuss some simple but surprisingly unexplored methods of bypassing restrictions.