[27307] in North American Network Operators' Group
Re: Does Anyone Care?
daemon@ATHENA.MIT.EDU (Brian Dickson)
Fri Feb 11 12:37:36 2000
Message-Id: <200002111733.MAA23259@spare.de.teleglobe.net>
From: "Brian Dickson" <briand@spare.de.teleglobe.net>
To: Omachonu Ogali <oogali@intranova.net>
Cc: Bjorn Carlsson <bc@ebone.net>, nanog@nanog.org, ras@above.net,
noc@qwest.net, noc@ebone.net, noc@sprint.net, briand@teleglobe.net
In-reply-to: oogali's message of "Fri, 11 Feb 2000 05:14:56 -0500".
<Pine.BSF.4.10.10002110513100.17721-100000@hydrant.intranova.net>
Date: Fri, 11 Feb 2000 12:33:04 -0500
Errors-To: owner-nanog-outgoing@merit.edu
> But why would the interface always transmit broadcasts to the FDDI
> interface in Spokane and show excessive latency at anytime of the day
> that you try to connect/send an echo reply at it.
A) Because *multicast*, when sent on a *broadcast* media like FDDI, is
sent as *broadcast* packets (to avoid replication);
B) Because multicast is very VIP2 CPU-intensive, doing all kinds of funky
things - we see that on very low utilization, there is high CPU. This in
turn shows up as latency *to* the router, but not *through* the router
(different switching paths).
Basically, this is normal, and ~5Mbps is about right for this traffic.
> On Fri, 11 Feb 2000, Bjorn Carlsson wrote:
>
> > Hi,
> >
> > Not sure I follow. schnell.ebone.net is actually an interface of Sprint
> > icm-bb1-pen which connects to a FDDI ring in Pennsauken built for multicast.
> > The name/address (schnell.ebone.net) is there for historial reasons and
> > should be changed to something-else.icp.net.
> >
> > As for directed broadcast it is long since turned off on all EBONE
> > routers.
> >
> > --BC
> >
--
Brian Dickson, Email: briand@teleglobe.net
Director, Backbone Engineering Tel : +1 703 755 2056
Teleglobe Communications Corp., Fax : +1 703 755 2648
Rm 3214, 11480 Commerce Park Drive, Cell : +1 703 851 9053
Reston, Virginia, USA, 20191 http://www.teleglobe.com
