[36266] in North American Network Operators' Group
Re: Sprint and peering points
daemon@ATHENA.MIT.EDU (Sean M. Doran)
Sun Apr 1 03:47:14 2001
To: deepak@ai.net, garlic@garlic.com
Cc: nanog@merit.edu
Message-Id: <20010401074228.17EC8C7901@cesium.clock.org>
Date: Sat, 31 Mar 2001 23:42:28 -0800 (PST)
From: smd@clock.org (Sean M. Doran)
Errors-To: owner-nanog-outgoing@merit.edu
garlic@garlic.com writes:
| So by making the change Sprint unilaterally shifts the transit packets from a
| public peering point away from themselves.
So if a *customer* asks Sprint, either directly, or via a
community like 1755:12xx (see whois -h whois.ripe.net AS1755),
is this "unilateral" on the part of Sprint?
| What would have been nice is for Sprint to tell its customers it was doing
| this. Then I would have expected the change in inbound traffic flows and
| taken action.
If it were a paying customer who indicated that Sprint should
prepend, do you feel Sprint should be obliged to inform all
customers beforehand? Is your answer different in the case
of non-revenue connections ("peers")?
How should this be done in the event that "prepend-request"
BGP communities are being used by a network connected to
Sprint, given that the other network may set or not set
the attribute for any given prefix at any given time?
| As it was, I opened a trouble report and wasted a lot of time looking
| for a problem.
Welcome to the Internet, it has routing complexity growth over time!
Sean.
