Re: Traffic shaping per connection

From: Charl Matthee (charl@mailgate.net)
Date: Fri Sep 15 2000 - 18:17:11 EST


On Fri, Sep 08, 2000 at 08:54:49AM +0100, Colin Watson wrote:

> > You realize of course that CBQs efficiency diminishes with each added
> > connection? Its not designed for a high volume of "classes"
>
> I do, yes. For the moment I'm not too bothered about it; I'm actively
> trying to slow things down, as it's the point of the exercise. However,
> if diminished efficiency means that I'm going to start seeing a lot of
> asymmetry in flow rates, then I might have to worry about it a bit more.

We were running a little solution that would dynamically
create/destroy classes as needed. We were running an average of
150-180 concurrent classes and CBQ did much like this.

Class specifications were disregarded, parent classes were overrun and
sibling classes did not share bandwidth according to specification.

> Will the fact that I'm organizing my classes in a binary tree speed
> lookup times, or am I barking up the wrong tree here?

I can't see how this will benefit you at all.

Ciao

Charl
__________________________________________________________________________

              The principle cause of problems is solutions.
__________________________________________________________________________

      Charl Matthee +27-11-721-3800
      Reality Manufacturing +27-11-405-6508
__________________________________________________________________________
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org



This archive was generated by hypermail 2b29 : Fri Sep 15 2000 - 21:00:29 EST