[122001] in North American Network Operators' Group
Re: How polluted is 1/8?
daemon@ATHENA.MIT.EDU (Joel M Snyder)
Wed Feb 3 15:10:49 2010
Date: Wed, 03 Feb 2010 13:10:07 -0700
From: Joel M Snyder <Joel.Snyder@Opus1.COM>
In-reply-to: <mailman.3296.1265225836.817.nanog@nanog.org>
To: nanog@nanog.org
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
> Having this data is useful, but I can't help to think it would be
> more useful if it were compared with 27/8, or other networks. Is
> this slightly worse, or significantly worse than other networks?
I have only anecdotal information regarding 45/8.
45/8 is assigned to Interop, and as such it is brought up-and-down as
Interop's shows move in and out of convention centers. Starting at
least 5 years ago, it has proved impractical to start announcing 45/8,
since this causes immediate and massive amounts of traffic to flow into
the show network.
The last time that I know that the full 45/8 was announced, traffic
settled down to about a full T3's worth of bandwidth before the network
engineers started announcing smaller /16 chunks as actually needed.
Even /16 has proved impractical while the network is being built-out,
before the show, because the build-out site typically has T1-ish
bandwidth---again, saturated with a /16 being announced.
This information is very different from the RIPE Labs experiment which I
think showed that certain "obvious" addresses (1.1.1.1 seemed to be the
kicker in my short reading of their report) were being mis-used heavily.
But I suspect that 27/8 would have similar issues to 45/8.
However, it is not clear to me that this is different from any other /8.
In other words, for those that have a /8, they probably DO have to put
up with a T3-worth of garbage flowing their way before they move the
first useful packet. However, you don't get a /8 unless a T3 is small
potatoes to you, hence...
jms
--
Joel M Snyder, 1404 East Lind Road, Tucson, AZ, 85719
Senior Partner, Opus One Phone: +1 520 324 0494
jms@Opus1.COM http://www.opus1.com/jms
