[69994] in North American Network Operators' Group
Re: Backbone IP network Economics - peering and transit
daemon@ATHENA.MIT.EDU (Deepak Jain)
Fri Apr 23 13:48:40 2004
Date: Fri, 23 Apr 2004 13:47:30 -0400
From: Deepak Jain <deepak@ai.net>
To: Michel Py <michel@arneill-py.sacramento.ca.us>
Cc: alex@yuriev.com, nanog@merit.edu
In-Reply-To: <DD7FE473A8C3C245ADA2A2FE1709D90B0DB0C9@server2003.arneill-py.sacramento.ca.us>
Errors-To: owner-nanog-outgoing@merit.edu
> If "direct connecting" != peering then definitely.
>
> Maybe we need to say differentiate between:
> - Connected transit
> - Remote transit
> - Connected peering
> - Remote peering
>
> And agree that, by default,
> transit ~= remote transit
> peering ~= direct peering
Without getting too complicated.
transit is always direct connection to a single AS, and indirect to all
others. For simplicity's sake, single-homed customer ASes behind the
transit AS are not considered apart from the transit AS. It is indirect
for the rest of the internet, including the sum of all peering (read:
direct connection without any indirect connections) connectivity.
peering is a always direction connection to a single AS and no indirect
connections are expected. Again, single-homed customer ASes are
considered part of the peering AS.
ASes that can only be reached from a single AS can only be reached by
those with a direction connection to the upstream AS.
---
This model [good or bad] allows people who pay for customer-only routes
from a transit provider they can't settlement-free peer with be
considered in the same breath as "true peers". For technology concerns,
I think this is valid. For business reasons there is probably some
difference.
DJ
