December 2016 - ooni-dev - lists.torproject.org

Procedure for reporting false positives
by Khairil Yusof 15 May '17

15 May '17

I'm reviewing the Malaysian Citizen Lab test url report summary and there seem to be false positives (or possible interference that is not visible) The ones that are obviously outdated, and redirected due to domain parking or something I file for removal (usually short lived minor porn sites) Where should I report and track these? http://www.yawa.org on-line gambling site is reported as possible http- diff

3 3

tcp_connect tests sanitizing wisctorbridge02 bridge's address
by David Fifield 15 Dec '16

15 Dec '16

I mentioned this in #ooni IRC and Arturo explained what was going on. I'm making a post so I have something to link to in code comments. The tcp_connect nettest is now reporting reachability of Tor Browser default bridges and directory authorities. One of these bridges, host:128.105.214.161 port:8080 nickname:wisctorbridge02 protocol:fte is being sanitized in published reports (its address is removed and replaced with its hashed fingerprint). For example, see https://s3.amazonaws.com/ooni-public/sanitised/2016-12-14/20161213T003954Z-… There is only one line that starts with {"test_keys":{"bridge_hashed_fingerprint":... For all the other targets, the "input" key looks something like "fte 128.105.214.161:8080", but for this one it is the bridge's hashed fingerprint "626ec5e8b39dbc1d1b853529cdc80d21bf4c20e1" (which is the correct hashed fingerprint of the wisctorbridge02 bridge). Arturo says this is happening as an effect of the data pipeline sanitization. For some reason, this one bridge is in the list of bridges to be sanitized. As a workaround, I'm looking for this hashed fingerprint and mapping it back to the bridge's address and port in code.

1 0

A first look at tor_bridge_reachability
by David Fifield 15 Dec '16

15 Dec '16

Since https://github.com/TheTorProject/ooni-probe/issues/652 was merged, there have been about five days of tcp_connect tests for Tor bridges and directory authorities. I've posted some preliminary code and analysis here: https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… Interesting results so far: * Probes from 21 countries have contributed: BR, CZ, FI, GB, GR, ID, IN, IS, IT, KR, MM, MX, NL, NO, PK, TH, TR, TW, UG, US, VN. * The probes in ID and TH are reporting spurious reachability from destinations on port 80 (the destinations are down but the probes report connection success). * The probes in KR and MX are blocking some of the directory authorities. Here is a big illustration, with one page per probe destination. Blue dots are successful connection, gray dots are timeouts, and other colors indicate other connection errors. The pages are sorted alphabetically by nickname, so at the top you will see the bridge Azadi. Azadi has been down for a few days, so its recent dots are gray. https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… These pages show spurious reachability from ID and TH to Azadi:80 (obfs4) and fdctorbridge:80 (fte). These two bridges are both down: Azadi:80 is having a temporary outage and fdctorbridge:80 was decommissioned a long time ago. But two probe sites report them as reachable. https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… The next listed pages show potential blocking of certain directory authorities from KR and MX. Blocked only in MX: dannenberg:80 (dir_port) dannenberg:443 (or_port) gabelmoo:80 (dir_port) gabelmoo:443 (or_port) maatuska:80 (dir_port) maatuska:443 (or_port) Blocked in KR and MX: dizum:80 (dir_port) dizum:443 (or_port) moria1:9101 (dir_port) moria1:9131 (or_port) tor26:80 (dir_port) tor26:443 (or_port) Additionally, moria1:9101 and moria1:9131 are blocked for the probe identified as "ooni-ZZ-AS0". https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016… https://people.torproject.org/~dcf/graphs/ooni_tor_bridge_reachability/2016…

1 0