Dummy moat distributor

Implement a dummy moat distributor to distribute the censorship map (bridgedb#40025 (closed)).

It requires three API endpoints:

• Get Circunvention Map. Provide the static json file with the list of countries.
• Get circunvention settings. Use the IP of the client (or a provided param) to return the right settings for the country.
• Get BuiltIn bridges.

As long as #12 (closed) is not ready we'll provide dummy bridges in the circunvention map to test that everything works properly.

changed milestone to %Sponsor 30 - Objective 3.3

added Doing Sponsor 30 - FINISHED labels

assigned to @meskio

marked this issue as related to bridgedb#40025 (closed)

marked this issue as related to #60 (closed)

changed time estimate to 8h

To discover the country from the IP I'm looking into options. On a fast look I found:

• geoip2. A go library that uses maxminddb and is used by many projects and looks maintained.
• snowflake implementation. The snowflake broker does implement it's own parser for the geoip DB provided by tor, that AFAIK is the ipfire (but the debian package's copyright claims to be something from webhosting.info that seems to be an expired domain).
• bridgedb geo.py. BridgeDB has it's own parser as well, but instead of using the tor's geoip DB uses maxminddb (from the geoip-database debian package).

I'm wondering what of the two DBs are better if ipfire or maxminddb, why every project uses it's own implementation.

I guess the two options I see are to export the snowflake implementation to our own tor-geoip parser library to be used by different projects or to use geoip2 library.

It looks like maxmind has changed the license recently and forced tor to find other options. I think the best will be to use the tor's provided geoip db, as tpo will keep it up to date and maintain the format between providers.

As a first step I'm going to copy code from snowflake to keep working on this. But I created a ticket to explore the option of a shared library (team#32 (closed))

added 4h of time spent

mentioned in issue team#32 (closed)

Basic implementation done: https://gitlab.torproject.org/meskio/rdsys/-/commits/moat

changed time estimate to 10h

added 6h of time spent

mentioned in merge request !15 (merged)

added Needs Review label and removed Doing label

After talking with @richard on irc we decided to do some changes to it:

• the request types field should be renamed transports
• rethink the distribution field, maybe ditch it and create another transport type called obfs4-built-in?

added Doing label and removed Needs Review label

added Needs Review label and removed Doing label

Let's produce errors in the same way than the current moat does. Returning something like:

{
  "errors": [
    {
      "code": xxx,
      "detail": "error string",
    }
  ]
}

Now it can produce two different error codes:

• 400 for malformed requests
• 406 for country not found

added Doing label and removed Needs Review label

changed time estimate to 12h

added 2h of time spent

added Needs Review label and removed Doing label

changed milestone to %Sponsor 96: Rapid Expansion of Access to the Uncensored Internet through Tor in China, Hong Kong, & Tibet

added Sponsor 96 label and removed Sponsor 30 - FINISHED label

mentioned in issue #69 (closed)

added Next label and removed Needs Review label

closed