The paper is a general description of Censored Planet, a censorship measurement platform that specializes in remote measurement techniques. Censored Planet aims for coverage that is both global and continuous, capable of supporting regular longitudinal measurements as well as focused, rapid studies. This paper illustrates every step of the pipeline: curating inputs, choosing vantage points, running scans, processing results, and checking for instances of censorship. There are subtleties throughout that are not immediately apparent but that have an effect on the quality of measurements and the trustworthiness of results.
Censored Planet combines multiple modular remote measurement techniques. Each has advantages and drawbacks. They complement the direct measurement techniques of platforms like OONI and ICLab, which are generally more flexible in the types of tests they can run, but are more limited in network and temporal coverage.
Augur – uses TCP and IP side channels to infer IP blocking between two remote endpoints.
Satellite/Iris – scans open DNS resolvers for anomalous responses.
Quack – uses the TCP echo and discard protocols to check for inbound and outbound pattern-based blocking.
Hyperquack – detects bidirectional blocking of HTTP and HTTPS.
Table 1 has a list of past censorship events that were automatically detected through time series analysis of Censored Planet data. Some of them had not previously been reported. The researchers provide commentary on censorship in Sri Lanka, Norway, Turkmenistan, Zimbabwe, and Japan. One of the trends they highlight is increasing censorship of HTTPS over their 20 months of measurements.
This paper is a kind of snapshot of Censored Planet as it existed in 2020. The platform has been continually refined since then, especially in how measurements are run and data are analyzed. You can read about more recent updates at https://docs.censoredplanet.org/.
Thanks to Ram Sundara Raman for comments on a draft of this summary.
Censored Planet: An Internet-wide, Longitudinal Censorship Observatory Ram Sundara Raman, Prerana Shenoy, Katharina Kohls, Roya Ensafi https://censorbib.nymity.ch/#Raman2020c https://censoredplanet.org/censoredplanet
The paper is a general description of Censored Planet, a censorship measurement platform that specializes in remote measurement techniques. Censored Planet aims for coverage that is both global and continuous, capable of supporting regular longitudinal measurements as well as focused, rapid studies. This paper illustrates every step of the pipeline: curating inputs, choosing vantage points, running scans, processing results, and checking for instances of censorship. There are subtleties throughout that are not immediately apparent but that have an effect on the quality of measurements and the trustworthiness of results.
Censored Planet combines multiple modular remote measurement techniques. Each has advantages and drawbacks. They complement the direct measurement techniques of platforms like OONI and ICLab, which are generally more flexible in the types of tests they can run, but are more limited in network and temporal coverage.
Table 1 has a list of past censorship events that were automatically detected through time series analysis of Censored Planet data. Some of them had not previously been reported. The researchers provide commentary on censorship in Sri Lanka, Norway, Turkmenistan, Zimbabwe, and Japan. One of the trends they highlight is increasing censorship of HTTPS over their 20 months of measurements.
This paper is a kind of snapshot of Censored Planet as it existed in 2020. The platform has been continually refined since then, especially in how measurements are run and data are analyzed. You can read about more recent updates at https://docs.censoredplanet.org/.
Thanks to Ram Sundara Raman for comments on a draft of this summary.