[tor-talk] How accurate does need the clock to be?

Maxim Kammerer mk at dee.su
Mon Feb 20 16:46:57 UTC 2012


On Mon, Feb 20, 2012 at 18:08,  <proper at tormail.net> wrote:
> How accurate does the clock have to be for Tor's correct operation?

This is not a simple question. See:
https://tails.boum.org/contribute/design/Time_syncing/
https://mailman.boum.org/pipermail/tails-dev/2012-February/000878.html

My understanding is as follows: if your clock is m minutes behind (m <
60), then Tor won't operate for up to m minutes after fetching
consensus for the first time (m minutes if it fetches at round hour
correct time, less if later), but should operate fine afterwards. If
your clock is up to 3 hours forward, Tor should operate correctly.
Accessing hidden services has stricter requirements (±1 hour?), but
apparently that has been fixed in alpha Tor versions.

Tor developers, feel free to correct me.

I think that Tor could include a switch for ignoring clock correctness
when verifying consensus, but this is apparently an ideological issue.

-- 
Maxim Kammerer
Liberté Linux (discussion / support: http://dee.su/liberte-contribute)


More information about the tor-talk mailing list