[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-talk] Extend auto-IP-switching-time in TorBrowser (and depending from time of inactivity)



On Tue, Mar 22, 2016 at 05:08:35PM -0700, Mike Perry wrote:
> Ben Stover:
> > As far as I know TorBrowser switches automatically every 10 minutes the node chain resp. the IP of the ExitNode.
> > 
> > Can I somehow extend this timeout time to another value e.g. 30 minutes?
> > 
> > Or (even better) can I let Tor auto-switch the IP and chain depending from the time of inactivity (.e.g when 15 minutes no
> > web page is called)?
> 
> We had a long discussion about this in
> https://trac.torproject.org/projects/tor/ticket/15482. Ultimately, a fix
> was merged to Tor, but it did not cause Tor to update its circuit
> discard timeout (the "dirtyness" timeout) upon stream detach.
> 
> I have also noticed worse behavior since Tor Browser switched from the patch I
> wrote in
> https://trac.torproject.org/projects/tor/attachment/ticket/15482/0001-Bug-15482-Don-t-abandon-circuits-that-are-still-in-u.patch
> to the version in Tor today.
> 
> I also agree we should be more aggressive about keeping circuits in use.
> I think we should go back to updating this timeout when streams are
> closed, otherwise we risk the situation where HTTP KeepAlive keeps an
> idle stream open for several minutes, and then when that stream closes,
> it is more likely that a new stream will go on a separate circuit
> because the timeout expired while the stream was open but idle.

I'm confused. What is the situation you are concerned about?  A new
stream would go over a new circuit whether the previous stream is kept
alive or not. Is that not so? And I'm not sure what keeping the idle
stream open has to do with this. I understand the UX issues of
switching circuits, and I get the threat if not allowing a stream to
close causes attaching indefinitely to new circuits.  But I don't
understand why it is bad to have a new stream open on a new circuit
after another stream closes that was artificially kept open for a
while vs. having the new stream open after an initial stream closed
normally.

aloha,
Paul
-- 
tor-talk mailing list - tor-talk@xxxxxxxxxxxxxxxxxxxx
To unsubscribe or change other settings go to
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-talk