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

Re: Asynchronous bandwidth limiting



First off, thanks for running a node - the network always needs more bandwidth.

As far as i'm aware, it isn't possible to specify incoming and outgoing limits separately, and if it were, the outgoing would always be higher. For the most part, relayed traffic is pretty close to 1:1; for everything that comes in, there is equal data going out (to the next node in the chain, the source, or the destination). The one major exception to this is the directory; requests for the directory are very small, but the results can be pretty large - but, that just means more outgoing than incoming. There is no benefit to having more incoming bandwidth than you do outgoing bandwidth. You can always disable the directory (DirPort 0) if you want to avoid that little bit of outgoing traffic, but usually it isn't too significant.

Hope that helps.

 - John Brooks


On Tue, Jan 13, 2009 at 2:48 PM, Sebastian Lechte <seb@xxxxxxxxxxxxxxxx> wrote:
Hi everyone,


I plan to set up a tor node later this week and have been reading
documentation and different wikis. I have found how I can limit the
total bandwidth the server will consume.

My hoster only limits outgoing traffic, but as far as I understand, it
is not possible to specify bandwidth-limits separately for incoming and
outgoing traffic.

Would it be good to be able to specify limits for outgoing or incoming
traffic separately, or would that only complicate matters without
providing any benefit?
Do incoming and outgoing traffic/bandwidth even differ significantly?
Does this depend on the type of node and exit-policy (if any) one uses?

I have no experience and have not found specific information on this, so
I hope that the list members can provide information on this question or
provide directions for further reading.


Sincerely,

S.Lechte