Glad to help. Good luck in your configuration.
Heres a few websites about effective traffic shaping:
Also keep in mind that some people claim that inbound traffic shaping is
impossible/undesireable, while others use inbound traffic shaping and say
nothing but good things about it. You really need to experiment with it on
your own and make a decision for yourself.
----- Original Message -----
From: "Carlo" <lists at nethq dot org>
To: "Eric Shorkey" <eshorkey at commonpointservices dot com>
Cc: "Thomas Hertz" <term at cynisk dot net>; <m0n0wall at lists dot m0n0 dot ch>
Sent: Wednesday, June 09, 2004 7:56 AM
Subject: Re: [m0n0wall] Wondershaper
> Hello Eric,
> Eric Shorkey wrote:
> >He's already using the traffic shaper to limit his outbound smtp
> >That's not his question. He's asking if queue support was ever
> >The answer is yes. Queue support has been implemented for a while now.
> >not the easiest thing to set up though. I know the queue support is in as
> >early as version 1.0. To configure it, you need to create a pipe with the
> >maximum bandwidth you wish to use for all involved queues. (Usually
> >something just slightly less than your maximum throughput works well.)
> >create a queue for each service you wish to have share the pipe. Don't
> >forget to create a default queue that matches everything, otherwise
> >only be sharing bandwidth between matching services, and nothing else.
> >your weights as necessary. Remember that every queue is guaranteed
> >bandwidth, regardless of it's weight. How much bandwidth is determined by
> >its weight vs all the other weights added together. And NEVER set the
> >to anything higher than 0. Otherwise you'll end up with 2 problems. Your
> >packets will be held at the firewall for the value of delay in
> >and you'll end up with a very busy firewall when the traffic gets thick
> >you might start dropping packets. I really don't know why we even have
> >delay option in the gui, since all it does is create detrimental effects.
> Thank you for your brief explanation, i will try to resolve my issues
> with your advise.
> >Is all of this tough to set up? Not really, but you have to know what
> >doing first, otherwise you'll end up with a bad configuration and you'll
> >unexpected results. It's hard to really gauge where m0n0wall sits on the
> >ease of use vs configurability meter. I'd say it sits sort of in the
> >maybe leaning a little bit towards more configurability. Honestly, this
> >bad place to be. It's too hard for the people that just want to plug in a
> >box and have it work, and it's too lacking for people with demanding
> >My suggestion to you, Carlo, is go back to wondershaper. Seriously. I'm
> >trying to be mean, but if you don't already understand how to configure
> >iptables or ipfw, and don't understand the concepts behind packet
> >and already have a good understanding of ip routing, then m0n0wall is
> >to be too hard to configure correctly. The friendly looking interface to
> >m0n0wall makes it seem easier than it really is.
> I think my knowledge should be enough to administer m0n0wall on a decent
> level. It is not that i do not want to spend the time on it but that i
> cannot spend to much time into it. I think m0n0wall is a good solution
> for me because it is easy to administer over http with a good and clear
> interface. If i spend the time in reading/learning about traffic shaping
> i am sure i will be able to set it up correctly, but my thinking is that
> maybe the same or similar rule's could be already preconfigured and be
> enabled with a few basic settings just like how wondershaper works.
> Maybe for somebody like you (who knows far more about this subject then
> me) could create it and submit it to the mailing list? I think it would
> be a valuable addition to m0n0wall for unexperienced users.
> > Smoothwall might be a good
> >choice for you as well. It's pretty easy to set up, though I don't think
> >free version has any packet shaping support. You'd have to look to make
> I already purchase 4 soekris boxes, so switching to Smoothwall is out of
> the question.