[38797] in North American Network Operators' Group
RE: Multicast Traffic on Backbones
daemon@ATHENA.MIT.EDU (John Olp)
Thu Jun 14 23:34:01 2001
From: "John Olp" <john.olp@olpinc.net>
To: <nanog@merit.edu>
Date: Thu, 14 Jun 2001 20:35:50 -0700
Message-ID: <NEBBIJENBIMKECIBAAEOIECMCEAA.john.olp@olpinc.net>
MIME-Version: 1.0
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
In-Reply-To: <20010613153310.A6401@digex.net>
Errors-To: owner-nanog-outgoing@merit.edu
Sorry for getting in late on this but...
The argument that multicast should be billed based on the number of
receivers is flawed. Those receivers are already being billed based on the
bandwidth they use regardless of source.
In other words, the multicast source is not, as someone suggested, using
more backbone that it is paying for. Everyone receiving a multicast feed is
also picking up the cost of their respective traffic.
Bottom line is, for certain problems, multicast is simply much more
efficient than unicast. There are already killer apps, but there also seem
to be killer obstacles. Network providers are fully aware of the burdens of
unicast. But since they make more profit when their efficiency is up and
their customer's is not, they'll lean to keep that balance or try and charge
more for offsetting it. It will probably be a while before it is promoted
as the better solution.
-John
