Whonix Census

From Whonix
Jump to navigation Jump to search

Whonix User Count Statistic[edit]

Table: Whonix Project Metrics

Date Daily Whonix-Gateway Users
15 March 2019 3,682
24 August 2019 4,107
17 October 2019 4,446
8 February 2020 5,326
31 March 2020 5,887
1 March 2021 8,236

The user count is manually updated on occasion, but it is a low priority.

The daily Whonix-Gateway users estimate is impacted by several factors:

  • Total number of Whonix users: The total Whonix user population is probably higher than the number of daily Whonix-Gateway users.
  • New day, still running, new count: A Whonix-Gateway that runs longer than one day (24 hours) will re-fetch the Whonix warrant canary and therefore be counted again.
  • New day, new boot, new count: A Whonix-Gateway that has not been run for more than one day (24 hours) will re-fetch the Whonix warrant canary and therefore be counted again.
  • Duplication: Users relying upon multiple Whonix-Gateway are counted multiple times; this is difficult to avoid with the current design (see next section).
  • Non-duplication: Boots are not explicitly counted. Users counts that rely upon the number of Whonix-Gateway boots would lead to overestimates because many users start and shutdown their computers several times a day.
  • Non-Live: Live users are not counted.

Whonix User Count Design[edit]

There are strict requirements for the Whonix user count:

  • no transfer of private information
  • no unique identifiers are utilized
  • no collection of IP addresses [1]
  • count only once within a 24 hour period
  • provide an option to disable user counts
  • a by-product of, and based on systemcheckarchive.org's automated Warrant Canary Check [2]
  • downloaded over Tor using the Whonix v3 onion service [3]
  • secure operation [4]

The census design was inspired by Clean Insights (Privacy-Preserving Measurement)archive.org, [5] Matomoarchive.org, [6] Tor Metricsarchive.org, Tails boot statisticsarchive.org [7], Prioarchive.org [8] Privacy-Preserving Telemetryarchive.org by Mozilla. [9] and Divvi Uparchive.org. For further details, see: Warrant Canary Check.

Whonix User Count Rationale[edit]

A user count statistic provides several benefits:

  • Whonix maintainers can track if the user population is increasing over time.
  • Certain people or organizations decide whether to contribute based on the estimated impact of the Whonix project.
  • Statistics can help answer certain questions such as the estimated user population of different anonymity platforms.
  • Direct comparisons between the number of Tor, Tails and Whonix users become possible.
  • Users can make an informed decision about fingerprinting risks. [10]
  • User counts can inform broader Internet and network fingerprinting considerations.
  • The data is already available due to systemcheckarchive.org's automated verification of the Whonix warrant canary. [11]

Disable Whonix User Count[edit]

Refer to the following systemcheckarchive.org chapters:

Forum Discussion[edit]

See Also[edit]

Footnotes[edit]

  1. By using an onion service, the server has no way to log originating IP addresses. The whonix.org website privacy policy is different, see:
  2. Which verifies the Whonix warrant canary.
  3. This is similar to how sdwdate fetches time from onion time sources.
  4. This process has identical security to the Warrant Canary Check.
  5. Developed by the Guardian Projectarchive.org.
  6. Formerly Piwik, which is a free and open source web analytics application.
  7. Private, Robust, and Scalable Computation of Aggregate Statistics by Stanford University.
  8. Like whether they prefer sharing the same network fingerprint as thousands of other users (Whonix) versus millions of other users (Tor Browser in isolation).
  9. Previously Whonix News.

We believe security software like Whonix needs to remain open source and independent. Would you help sustain and grow the project? Learn more about our 12 year success story and maybe DONATE!