[10689] in Commercialization & Privatization of the Internet
Re: clarifying earlier NAP
daemon@ATHENA.MIT.EDU (Dan Lynch)
Sat Mar 5 14:36:38 1994
Date: 4 Mar 1994 20:38:35 -0800
From: "Dan Lynch" <dlynch@interop.com>
To: "Gordon Cook" <cook@path.net>, "William Manning" <bmanning@is.rice.edu>
Cc: com-priv@psi.com
Reply to: RE>>clarifying earlier NAP d
Bill, Thanks for making the issues so clear. You are very good at that.
And... The Anwser Is!!!::: NONE OF THE ABOVE. Not long term, that is. In
the short run there may be an advantage or two for going with one provider
model or the other, but in the long run (less than two years) there will be
enough money in the game (of Internetting) that "price, performance and
delivery" will be all that people look at. Because there will be enough
competition going on that customers can shop like they always want(ed) to.
Along any combinations of the above three parameters they choose.
Dan
--------------------------------------
Date: 3/4/94 7:02 PM
To: Dan Lynch
>From: William Manning
Gordon Cook sez: (in some time warp)
> UUNET & PSI....
> The Washington NAP, where if they did connect, they'd be most likely to,
> was not on the list of mandatory NAPs. I am not sure whether there will
> be policy differences in how Washington NAP is handled because it is not
> one of the three NAPs required by the solicitation.
Well, the do connect there and if I read Steves note & the solicitation,
the WDC nap is mandatory. The sol. says there can be more @ NSF's choice.
Policy should be the same.
> Here's a new question:
> SPRINT & IRC $$....
> If they are not, is there any overwhelming reason why they should
> interconnect at any NAP other than the NYC one that they are running?
>
> What if Steve Wolff gave us NAPs and no one came?
>
Buzzt. Thats two.
First question. No.... unless there were business reasons.
Second question. This would mean that the CIX is the right model -OR-
that commercial services were actually able to a) get it right, (some
have a very good clue and some are clueless) and b) agree to peer w/
each other at a mutual exchange point. (something they are not quite
as good at yet) The NAP concept would exist, but there would be a lot
more of them with many fewer connects. Steve has lent NSF weight to
a small number with lots of connects, which is actually the better
approch.
NAPs are a good thing. It is the level playing field... as long as there
is still a neutral Routing entity that does not play in the transit game.
Now one for you.
Which is the better deal, a CIX connection or a PAC-NAP connection and why?
--
Regards,
Bill Manning
------------------ RFC822 Header Follows ------------------
Received: by interop.com with SMTP;4 Mar 1994 19:02:34 -0800
Received: from lists.psi.com by polaris.interop.com (4.1/SMI-4.1)
id AA18661; Fri, 4 Mar 94 19:03:06 PST
Received: by lists.psi.com (4.1/SMI-4.1.3-PSI)
id AA20449; Thu, 3 Mar 94 22:25:43 EST
Return-Path: <bmanning@is.rice.edu>
Received: from psi.com by lists.psi.com (4.1/SMI-4.1.3-PSI)
id AA20344; Thu, 3 Mar 94 22:25:03 EST
Received: from brazos.is.rice.edu by psi.com (4.1/2.1-PSI/PSINet)
id AA05613; Thu, 3 Mar 94 22:25:18 EST
Received: by brazos.is.rice.edu (AA25548); Thu, 3 Mar 94 21:25:06 CST
>From: bmanning@is.rice.edu (William Manning)
Message-Id: <9403040325.AA25548@brazos.is.rice.edu>
Subject: Re: clarifying earlier NAP discussion with Marvin
To: cook@path.net (Gordon Cook)
Date: Thu, 3 Mar 1994 21:25:06 -0600 (CST)
Cc: com-priv@psi.com
In-Reply-To: <9403021430.aa10262@pandora.sf.ca.us> from "Gordon Cook" at Mar
2, 94 02:30:04 pm
X-Mailer: ELM [version 2.4 PL22]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1618
