Is "gatereloaded" a Bad Exit?

Gregory Maxwell gmaxwell at
Mon Jan 31 17:47:37 UTC 2011

On Mon, Jan 31, 2011 at 11:30 AM, Andrew Lewman <andrew at> wrote:
> If we're going to start censoring Tor exits based on impressions, we
> might as well start blocking Tor relays that are rumoured to be run by
> national intelligence agencies, criminal organizations, martians, and
> other people we might not like.  In fact, we might as well go back to
> the original model of "every Tor relay operator has met and gained
> Roger's trust".

I'd disappointed that you're not responding to the argument I initially posed.

We should do this not because the node looks suspicious but because we
want to shape the behaviour of exit operators.
There are legitimate reasons why tor supports an operator controlled
exit policy,  but no real suggestion has been made for a _legitimate_
reason to allow 80 and block 443.

So we make carrying 443 part of the price of being a port 80 exit.  So
this exclusion shouldn't be seen as something that will eliminate bad
guys,  — it clearly won't — but will instead force them to behave more
like we want them to, by adding 443 capacity and making tor faster for
https users— as part of the price of operating an exit.

The best argument against this would be that it makes it harder for
people to spot these probably-bad nodes based on the exit policy and
exclude them for themselves.  I think this downside is inconsequential
because almost no one is actually going to do this, and anyone that
sophisticated can be trusted to protect themselves in other ways.

Tor has a great many behaviour shaping incentives in the protocol and
implementation.  This would not stand out as too unusual.
To unsubscribe, send an e-mail to majordomo at with
unsubscribe or-talk    in the body.

More information about the tor-talk mailing list