[tor-relays] Rampup speed of Exit relay

[D. S. Ljungmark]

On ons, 2016-09-21 at 17:29 +0200, Aeris wrote:
> > 
> > 17 MBytes/s in each direction.
> 
> From Atlas graph, your node is currently growing up, so wait few
> weeks more to 
> have the real bandwidth consumption, but don’t expect huge change.

It looks as if it stabilized a while ago, and I see more fluctuations
in between hours of the day, than I see between days.

> 
> 17M*B*ps is 140M*b*ps and you already have a good relay :)
> This is around the speed expected for standard CPU (150 to 300Mbps
> per Tor 
> instance, best CPU available can "only" drain around 500Mbps).


For a relay, yes. However, on an Exit node, the traffic patterns are
quite different than a pure relay, that's why I'm asking here for that.

Exit traffic uses a metric ton more, short-lived connections to
different parts of the world, and there's a lot of things there that I
could probably tune better.  ( Normal relays seldom have a trouble
filling the conntrac tables, while exits do. )

> 
> And your CPU have all chance to be the bottleneck at this speed. Tor
> is not 
> multi-core at the moment and so you can’t be able to fully use your
> CPU 
> capacity. For example, if you have a 4-core CPU, don’t expect to have
> more 
> than 0.25-0.3 load with only Tor (1 core fully used).

Load is only about IO, processes waiting for something. That's why I
quoted CPU usage, and the cpu usage per cpu isn't quite getting there
for me. :-(

> 
> You have to start another Tor instance to use a little more your CPU
> (1 other 
> core) and so to drain additionnal 150-300Mbps.


Scaling up on more hardware is always an option, but I really want to
push the limit of the exit node, as the others won't be exits (Local
network design, really)	, and exit traffic is always more
interesting.

//D.S.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://lists.torproject.org/pipermail/tor-relays/attachments/20160922/c644cf19/attachment.sig>