[tor-bugs] #5040 [Obfsproxy]: Make public bridges add obfsproxy stats to their extra-info descriptors

Tor Bug Tracker & Wiki torproject-admin at torproject.org
Thu Mar 15 07:57:31 UTC 2012


#5040: Make public bridges add obfsproxy stats to their extra-info descriptors
-------------------------+--------------------------------------------------
 Reporter:  karsten      |          Owner:  asn
     Type:  enhancement  |         Status:  new
 Priority:  normal       |      Milestone:     
Component:  Obfsproxy    |        Version:     
 Keywords:               |         Parent:     
   Points:               |   Actualpoints:     
-------------------------+--------------------------------------------------

Comment(by karsten):

 Replying to [comment:1 asn]:
 > Do we care about bridges giving out separate statistics for "normal"
 usage, and for pluggable transport usage? Can this harm anonymity? If we
 like doing this, we should spec. it.

 A fine question.  I'd say let's do this similar to how we're planning to
 extend IPv6 bridge usage statistics: 1) we count all connecting bridge
 users by country, regardless of IPv4 or IPv6, and 2) we add a new
 statistics how many users are IPv4 users and how many are IPv6 users.  See
 #5055.

 In the context of obfsproxy statistics, 1) is what you describe:

 > The implementation plan for providing client geoip statistics to bridge
 authorities could be something like: When using Extended ORPort, everytime
 we receive a client address through `USERADDR`, pass it to
 `geoip_note_client_seen()` like we are currently doing in
 `connection_or_set_state_open()`.

 For 2), I could imagine a stats line like this:

 {{{
 bridge-transport or=NN,obfs2=NN,...
 }}}

 The `NN` parts would be unique addresses connecting via a given transport.
 These sets could overlap if a user connects via multiple transports, but
 that would be fine.  The stats would tell us when a given transport sees
 increasing/decreasing use.

 I think 2) is somewhat harder to implement in geoip.c, because we'd have
 to add transport information to the data structure holding unique IP
 addresses.

 What these stats would not tell us is when a given transport stops working
 in a given country.  That's similar how the IPvX stats won't tell us when
 a given country blocks a given IP version.  I ''think'' that's okay.
 Other thoughts?

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/5040#comment:2>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online


More information about the tor-bugs mailing list