Hello,
The OONI team warmly welcomes you to join us next Tuesday for our monthly
community meeting.
*=> Where? *OONI Slack channel: https://slack.ooni.org/ (bridged with IRC:
ircs://irc.oftc.net:6697/#ooni)
*=> When?* Tuesday, *30th November 2021 at 14:00 UTC* (for 1 hour)
Please add topics that you'd like to discuss during the meeting in this
pad: https://pad.riseup.net/p/ooni-community-meeting-keep
The monthly OONI community meetings aim to:
* Collect community feedback on OONI tools & methodologies
* Address questions in relation to the use of OONI tools and OONI data
* Foster discussions on internet censorship issues
* Receive updates from the community
We hope you can join us!
All the best,
~ OONI team
Hello,
A few months ago, the OONI team collaborated with researchers at *IODA,
Kentik, UC San Diego, and University of Michigan / Censored Planet* on a
research paper examining censorship events in Myanmar (through diverse
measurement datasets) following the military coup.
The paper, titled "*A multi-perspective view of Internet censorship in
Myanmar*", was published by FOCI 2021 here:
https://dl.acm.org/doi/pdf/10.1145/3473604.3474562
Today, we published a blog post which *summarizes the research paper
findings*, and which links to IODA's amazing YouTube video (discussing the
findings):
https://ooni.org/post/2021-multiperspective-view-internet-censorship-myanma…
Best,
Maria.
Hello,
In collaboration with IODA, we published a *new research report* which
analyzes *Mozilla telemetry* to investigate *internet shutdowns in Myanmar,
Uganda, Belarus, and Iran *over the last year.
Read our research report here:
https://ooni.org/post/2021-investigating-internet-shutdowns-mozilla-telemet…
The folks from Mozilla also published a blog post about our report here:
https://blog.mozilla.org/data/2021/11/08/detecting-internet-outages-with-mo…
Below we share information about the research and some key findings.
*# About the research*
Given how widespread Firefox usage is worldwide, could Mozilla telemetry be
a valuable resource for the Internet freedom community to investigate
Internet shutdowns?
To explore this question, we analyzed an aggregated dataset of network
activity based on Firefox web browser usage worldwide (access to which was
provided to us by Mozilla).
To evaluate how useful Mozilla telemetry is for researching Internet
shutdowns, we:
* Selected case studies (known shutdown events from January 2020 onwards)
* Analyzed Mozilla telemetry
* Compared Mozilla telemetry with other public datasets
We checked whether Mozilla telemetry provides signals for:
* Myanmar -> Internet outages following February 2021 military coup
* Uganda -> Nationwide internet outage amid 2021 elections
* Belarus -> Internet outages amid 2020 elections
* Iran -> Internet outages following 2020 elections
*# Summary of key findings*
We provide detailed analysis and findings for Myanmar, Uganda, Belarus, and
Iran in our report:
https://ooni.org/post/2021-investigating-internet-shutdowns-mozilla-telemet…
Overall, based on our analysis for these 4 countries, we found:
1) *Mozilla telemetry provides strong signals for high impact internet
shutdowns* -> When access to the internet was shut down entirely (such as
in Uganda and Myanmar earlier this year), we observed a complete absence of
Mozilla telemetry. This corroborates the absence of metrics that we also
observe in other public datasets (such as IODA, Google traffic data, etc.)
that are used to investigate internet shutdowns.
2) *Mozilla telemetry can also be used to infer lower impact internet
shutdowns* -> When an internet shutdown is not "total" (when an internet
shutdown doesn't impact all networks and/or regions in a country), it can
still be possible to potentially infer that there is some internet
disruption from the high percentage of connection timeouts and unreachable
connections that are visible through Mozilla telemetry. We observed this,
for example, for "lower impact" internet shutdowns in Belarus and Myanmar.
Of course, it's important to cross-reference this data with other public
datasets (such as IODA and Google traffic data).
3) *The geographical granularity of Mozilla telemetry provides new research
opportunities* -> In Belarus, Mozilla telemetry showed a spike in
connection timeouts and unreachable connections from multiple locations,
suggesting that there may have been some reshaping of the network topology
during the August 2020 Internet outage.
4) *Mozilla telemetry does not seem to provide signals for short-lived
internet shutdowns (lasting less than 1 hour)* -> In Iran, Mozilla
telemetry did not provide signals of the two short-lived internet outages
that occurred on 3rd and 11th March 2020 (both of which lasted for less
than an hour, and which were visible through IODA data). This is likely due
to the fact that Mozilla telemetry is aggregated in hourly buckets,
therefore missing short-lived internet outages.
Overall, our analysis demonstrates that *Mozilla telemetry is a valuable
resource for investigating Internet shutdowns worldwide*, providing novel
insights through geographical data granularity.
We encourage Mozilla to *publish aggregated Mozilla telemetry as open data*
to support research & advocacy efforts investigating Internet shutdowns
worldwide.
We also share further recommendations in our report:
https://ooni.org/post/2021-investigating-internet-shutdowns-mozilla-telemet…
We thank Mozilla for providing us access to Mozilla telemetry for this
research, and for considering our recommendations.
Best,
OONI team.