[LARTC] How much does it matter where the throttling is done?

[LARTC] How much does it matter where the throttling is done?

[Brian Capouch]

I got into a "spirited discussion" tonight about just what the upstream 
effects of traffic shaping look like.

In the case in point, a private WAN with quite a few routers connects to 
a number of Internet POPs.  In some cases, the "leaves" are three or 
four hops from the backbone source.

The central focus of the discussion was the relative harm/benefit from 
putting the traffic shaper at the point where the bandwidth hits the 
Internet, at one extreme, versus at the "last-hop" routers on the other.

I won't go into the details as I suspect there is probably a cut and 
dried answer.

And rather than proffering my own ideas, I would rather not embarrass 
myself and just ask the experts.

Thanks.

B.

_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

Re: [LARTC] How much does it matter where the throttling is done?

[Stef Coene]

On Monday 25 November 2002 07:25, Brian Capouch wrote:
> I got into a "spirited discussion" tonight about just what the upstream
> effects of traffic shaping look like.
>
> In the case in point, a private WAN with quite a few routers connects to
> a number of Internet POPs.  In some cases, the "leaves" are three or
> four hops from the backbone source.
>
> The central focus of the discussion was the relative harm/benefit from
> putting the traffic shaper at the point where the bandwidth hits the
> Internet, at one extreme, versus at the "last-hop" routers on the other.
>
> I won't go into the details as I suspect there is probably a cut and
> dried answer.
>
> And rather than proffering my own ideas, I would rather not embarrass
> myself and just ask the experts.
It all depends on the network configuration.  You have to shape on the 
bottleneck on the network.  If you don't do that, your shaping can be undone 
by the bottleneck.  

Stef

-- 

stef.coene@docum.org
 "Using Linux as bandwidth manager"
     http://www.docum.org/
     #lartc @ irc.oftc.net

_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

Re: [LARTC] How much does it matter where the throttling is

[Art Reisman]

On Monday 25 November 2002 07:25, Brian Capouch wrote:

I agree with Stef, where is the bottleneck? Not only
that but what are you trying to shape? Without knowing
the other constraints and variables, by debating
somebody who holds different variables as constraints
you are sure to always get difference of opinion.

Art Reisman
www.apconnections.net



> I got into a "spirited discussion" tonight about
just what the 
upstream
> effects of traffic shaping look like.
>
> In the case in point, a private WAN with quite a few
routers connects 
to
> a number of Internet POPs.  In some cases, the
"leaves" are three or
> four hops from the backbone source.
>
> The central focus of the discussion was the relative
harm/benefit 
from
> putting the traffic shaper at the point where the
bandwidth hits the
> Internet, at one extreme, versus at the "last-hop"
routers on the 
other.
>
> I won't go into the details as I suspect there is
probably a cut and
> dried answer.
>
> And rather than proffering my own ideas, I would
rather not embarrass
> myself and just ask the experts.
It all depends on the network configuration.  You have
to shape on 
the 
bottleneck on the network.  If you don't do that, your
shaping can be 
undone 
by the bottleneck.  

Stef

_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/