Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 11 Feb 1999 16:24:35 -0800 (PST)
From:      Julian Elischer <julian@whistle.com>
To:        Deepwell Internet <freebsd@deepwell.com>
Cc:        Dennis <dennis@etinc.com>, freebsd-isp@FreeBSD.ORG
Subject:   Re: Bandwidth limiting/trafic shaping
Message-ID:  <Pine.BSF.3.95.990211161537.5795A-100000@current1.whistle.com>
In-Reply-To: <4.1.19990211125823.00b7c340@mail1.dcomm.net>

next in thread | previous in thread | raw e-mail | index | archive | help


On Thu, 11 Feb 1999, Deepwell Internet wrote:

> Although kernel throttling is a good solution, I disagree with your
> thoughts that any of the "fancier" shaping mechanisms give you no
> additional features.  
> 
> Kernel throttling will never give you the relational shaping that
> class-based-queueing gives.  We have quite a few machines under a
> class-based queueing machine and have it tweaked quite well.  With class
> based queueing you can define "classes" of computers and define which
> machines have priority over others.  You can also give people "guaranteed
> information rates" and let them burst above into any unused space.  We also
> have ICMP shaped into a seperate class.  Most of these can't be
> accomplished with kernel throttling.

Who says that kernel throttling can't be class based? (e.g. CBQ/ALTQ)
Of course Dennis says "Physical limiter", which could be a number of 
things. He is however motivated by the ultimate motivator (food on the
table) to make sure it works well.

julian

> 
> Thanks!
> Terry Ewing
> 
> 
> At 02:47 PM 2/11/99 -0500, you wrote:
> >At 09:38 AM 2/11/99 +0200, you wrote:
> >>Hello everybody!
> >>
> >>Can someone comment about comparison of bandwidth limiting software like
> >>dummynet or bwmgr from ET inc. and alternative queuing schemes like ALQ
> >>with Class Based Queuing (CBQ)? All seem to provide similar effects but
> >>which is preferable in what situation?
> >>
> >>If I'll get enough feedback I'll post summary.
> >
> >There are a lot of fancy names out there, but there is no evidence that the 
> >fanciest ones work any better. Ours is a physical limiter, kernel based with 
> >no additional interrupt overhead and can handle just about any level of
> >traffic 
> >that the machine can handle without it. Our new HTML interface makes it 
> >easy to manage also.
> >
> >Others are free and may work just fine for you as well. :-)
> >


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95.990211161537.5795A-100000>