This is the supporting web page for a research project to examine how quickly censorship circumvention proxies (Tor bridges) are noticed and blocked by censors after being made public. We conducted the research by active measurement of proxies from the United States, China, Iran, and Kazakhstan.

David Fifield, Lynn Tsai, Qi Zhong
University of California, Berkeley
(Authors are listed alphabetically.)

Writing

Censors’ Delay in Blocking Circumvention Proxies
David Fifield and Lynn Tsai.
In Proc. of Free and Open Communications on the Internet, 2016, USENIX.
BibTeX.
Detecting Censor Detection
David Fifield, Lynn Tsai, Qi Zhong.
arXiv preprint, 2017-08-13.
Time delays in censors’ reactions (thesis chapter)
David Fifield.
2017-12-15.

Code & Data

Probe source code and data

git clone https://www.bamsoftware.com/git/repo.eecs.berkeley.edu/proxy-probe.git

Paper source code (includes graphing code)

git clone https://www.bamsoftware.com/git/repo.eecs.berkeley.edu/proxy-probe-paper.git

Dates of notable bridge-related events

significant_dates.txt (14 KB)

Kazakhstan code and data (more intensive, full obfs4 bootstrap tests)

kz-data/

The probe sites in probe.csv are identified by code name.

China 1 = bauxite + spline (two IP addresses in the same AS)
China 2 = mistletoe
Iran = leopard
U.S. = eecs-login

Backup of data at archive.org: proxy-probe-data.

Censors’ Delay in Blocking Circumvention Proxies

Writing

Code & Data