[8619] in Commercialization & Privatization of the Internet
NSF NAPs & CIX & CoREN
daemon@ATHENA.MIT.EDU (Erik E. Fair" (Your Friendly Postm)
Fri Nov 26 21:39:29 1993
From: "Erik E. Fair" (Your Friendly Postmaster) <fair@apple.com>
In-Reply-To: <9311251307.AA05422@brazos.is.rice.edu>
To: bmanning@is.rice.edu (William Manning)
Cc: com-priv@psi.com
Date: Fri, 26 Nov 93 18:38:54 -0800
Well, if the new NAPs are not going to have any AUP, then there's a
very simple win-win case, given what I've read here:
The contract (or "cooperative agreement") for NSF NAP is let to the
CIX, and exactly one NAP is established (the existing CIX), under the
CIX's existing rules ("everyone must route to all their customers, and
no settlements").
That way, we have no Inter-NAP transit problems, and no connectivity
problems. Of course, that does leave us with a "so how do I get my
network to the NAP" problem, but that's solved, right?
If you have more than one NAP, you *must* have a network that plays
transit for all comers between the NAPs. The IP routing model really,
really wants to have a "top of the tree that knows all, sees all" (i.e.
a backbone)so long as we keep on using "default" routes every place
beyond stub sites.
Aside from awarding the NAP stuff to the CIX or ANS, there is one other
possibility: CoREN. However, I find it quite distressing that none of
the founding members of the CIX were listed in the CoREN press release
some months ago. It has also been very quiet since then - no word about
CoREN at all.
Erik E. Fair apple!fair fair@apple.com
