[tor-dev] Scaling tor for a global population

Sebastian G. <bastik.tor> bastik.tor at googlemail.com
Wed Oct 1 17:07:30 UTC 2014

30.09.2014, 01:12 isis:
> isis [mash-up]
> [3]: Please, don't give all the shit relays to me as bridges. I think it's
>      less important scalability-wise (right now) to have a strict cutoff rate
>      for bridges, but eventually, when/if we ever have Bridge Bandwidth
>      Authorities, BridgeDB should cut off anything below some completely
>      arbitrary rate, like 100 KB/s. I've gotten a bridge (from
>      http://bridges.torproject.org) which was 28 B/s. Yes, *bytes*. That thing
>      was probably slowing down the rest of the Tor network just by *existing*
>      via its molasses-speeds blocking the Exit from continuing the response
>      after SENDME number of cells, which is probably eventually going to cause
>      TCP timeouts on the Exit's side and a whole bunch of other messes.

I think the Tor Project requires a high number of bridges to make
collection of all addresses harder for some adversaries. I'm aware that
adversaries can outrun the number of brides. This might or might not be
valid until you shut down all vanilla bridges.

Obviously bridges that don't provide too less bandwidth should not take
part in the network.

>>   What I usually recommend is to users is based on their bandwidth and how
>> frequently their IP changes.  If their connection is fast and their IP never
>> changes (eg, a desktop or server), then run a non-exit relay [2].  For a
>> laptop that moves to-from work, then a relay or bridge.
> Actually, anything with a constantly changing IP is a terrible idea for a
> bridge.
> Think of it this way: BridgeDB hands you a bridge line on Monday. On Tuesday,
> the bridge's IP rotates to something else. Now your Tor doesn't work. Wah-wah,
> sucks for you!

The bridge being down is indeed a problem. I got the same recommendation
of running a bridge on my home-network, rather than a relay (or exit)
from a person working for/on Tor. I passed that recommendation around,
IIRC even on this list, and no one complained.

It is problematic to have this information around and you, isis,
speaking up soooo late.

The changing IP part is beneficial for censorship circumvention. It
would be broken if clients could learn the new IP address automatically.

Distribution of bridges just sucks for circumvention of censorship. (Not
your work, isis.) That could be improved by handing out puzzles.

>> There's one issue if you remove all the small relays, only relays run by
>> the NSA will be around. Not many people have access to multi-megabit upload
>> speeds. And those that do might also be using bittorrent.
> I'm quite certain that I'm definitely not the NSA, and I run a multi megabyte
> exit relay[.]

I don't think you are the Tor Project either, isis. You are working for
the Tor Project. Well depending on the view all employees are the Tor

I'm not suggesting you would run those nodes for the NSA or that you
work for them in any way.

Thank you all for your work in upscaling Tor.

Sebastian G. bastik

More information about the tor-dev mailing list