[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [tor-bugs] #16389 [Tor]: Redesign the HS client descriptor cache
#16389: Redesign the HS client descriptor cache
-------------------------+-------------------------------------------------
Reporter: dgoulet | Owner:
Type: defect | Status: reopened
Priority: major | Milestone: Tor: 0.2.7.x-final
Component: Tor | Version:
Resolution: | Keywords: SponsorR, tor-hs, TorCoreTeam201508
Actual Points: | Parent ID: #16381
Points: |
-------------------------+-------------------------------------------------
Comment (by arma):
Did everybody just assume I picked the perfect number "5 minutes", or has
anybody thought through the time period required here?
Basically, any time interval is suboptimal, in that waiting a long time
will result in more instances of unhappy users who can't reach their HS,
and waiting a short time will result in duplicate attempts.
It seems like we want a nonce in the hidden service descriptor, and in the
cache we tag which descriptor we had when we tried it, and then expire --
no, that wouldn't work either, since the hidden service doesn't need to
publish a new version of its descriptor if it's just reusing all its old
intro points. So there isn't any explicit signal by the HS that it was
gone for a while and has now returned so maybe you want to retry.
In sum: with this patch are we still going to have cases where Alice tries
to visit a hidden service, finds it unavailable, then the HS goes back
online and tells her to try again, and for five minutes she fails?
Maybe this argues for hidden services putting the nonce into their HS
descs anyway, and publishing an updated HS desc every time they lose
contact with their intro points, to give clients a way to recognize when
the HS has acknowledged that things changed? Yuck.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/16389#comment:16>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs