[tor-bugs] #33815 [Core Tor/Tor]: vanguards with meek - do or don't?

[Tor Bug Tracker & Wiki]

#33815: vanguards with meek - do or don't?
--------------------------+----------------------------------
 Reporter:  cypherpunks   |          Owner:  (none)
     Type:  task          |         Status:  closed
 Priority:  Medium        |      Milestone:  Tor: unspecified
Component:  Core Tor/Tor  |        Version:
 Severity:  Normal        |     Resolution:  duplicate
 Keywords:  vanguards     |  Actual Points:
Parent ID:                |         Points:
 Reviewer:                |        Sponsor:
--------------------------+----------------------------------

Comment (by mikeperry):

 Q4) Btw, here's my best guess of the kind of defense that *would* work in
 the general case. At a minimum, services will need to have a traffic re-
 shaper that makes their traffic patterns look like the reverse of what
 they are - HTTP is an aymmetric protocol in that requests are typically
 much smaller than responses. So some kind of traffic shaping to reverse
 this asymmetry is necessary (see ALPaCA for an example:
 https://www.freehaven.net/anonbib/cache/applicationlayer-pets2017.pdf).
 Then some amount of cover traffic would need to be carefully added onto
 this shaped application layer traffic, too. And even then, high volume
 onion services will at best look like high volume web
 crawlers/scrapers/bots, not real users.

 I don't think that this is within reach of service operators. It is still
 an open research problem. See also
 https://github.com/torproject/tor/blob/master/doc/HACKING/CircuitPaddingDevelopment.md#14
 -other-deployment-constraints and the rest of that doc (which is the place
 for such info, rather than the vanguards doc).

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/33815#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online