[tor-relays] Options for Managing Relay Load (was: Re: Really strange)

[teor]

> On 4 Jan 2018, at 09:52, r1610091651 <r1610091651 at telenet.be> wrote:
> 
> Hi teor
> 
> Thanks for the reply. I'm not having issues with my relay, and I've seen the mail about extra users indeed. Just an observation, as I've just noticed few waves of these in last couple of hours. The start very abruptly and end just as well.
> Not an expert, but I find it statistically significant to receive so many connections (499 for one /24) form a single subnet to one specific middle node.
> Unless that can be explained.

If they are connecting to middle nodes directly, and handling consensus
weights badly, they could be misconfigured or experimental Tor clients.
Or there are bugs in some Tor versions that cause small guard weights,
even for non-guards. (Or they could be Tor2web or Single Onion Service
clients.)

A consensus weight of 1000 gets you a probability of about 0.003%.
So if there are 1 million new clients, ignoring the Guard flag, we would
expect 30*N of them to connect to your relay, where N is the number of
entry nodes each client chooses.

Recent clients that have stable connections have 2 active guards.
Older clients may have 3 active guards.
A client that fails lots of connections has a large number of active guards.
(Recent clients try to limit this number. Older clients do not.)

So having 17 active guards is not unreasonable.
(It's a bug in Tor, but it's not unreasonable.)
This would lead to you seeing ~500 connections at your relay.

T

--
Tim Wilson-Brown (teor)

teor2345 at gmail dot com
PGP C855 6CED 5D90 A0C5 29F6 4D43 450C BA7F 968F 094B
ricochet:ekmygaiu4rzgsk6n
xmpp: teor at torproject dot org
------------------------------------------------------------------------



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: Message signed with OpenPGP
URL: <http://lists.torproject.org/pipermail/tor-relays/attachments/20180104/0b18d60f/attachment.sig>