[1870] in Commercialization & Privatization of the Internet
Al Weis' Responses to My Questions about the ANS Agreements
daemon@ATHENA.MIT.EDU (Gordon Cook)
Sat Jan 4 00:17:18 1992
To: com-priv@psi.com
Date: 3 Jan 92 23:53:51 EST (Fri)
From: cook@tmn.com (Gordon Cook)
<<MESSAGE from>> Gordon Cook 03-JAN-92 23:53
cook@tmn
SUBJECT: Re: ANS Connectivity Agreement
MESSAGE from =uunet!ans.net!weis 30-DEC-91 20:42
> Yes, Gordon, we still have the Gateway and Cooperative Agreements.
The Connectivity Agreement states that a network is willing to accept
commercial traffic and let it flow across its network - - and these
networks are eligible for funds from the infrastructure pool.
>
> A Gateway Agreement says that a network will host both commercial and
research/education institutions. A Cooperative Agreement states that,
while a network is willing to carry commercial traffic, it doesn't want to
sell the commercial services. Instead, it contracts with ANS CO+RE to
sell these services, and it provides the connection between the ANS
network and its customers. Gateway and Enterprise Agreements fund the
infrastructure pool.
>
> I will be forwarding to you the Hitchhiker's Guide to the ANS
Agreements, which was distributed and discussed at the fall FARnet
meeting, for your reference. If you still have questions, give me a call.
>
> Al
>
> ---
>
> <<REPLY from>> Gordon Cook 17-DEC-91 12:56
> cook@tmn
> Thank you Al. I appreciate your taking the time to respond. I hope
the Hitchhikers Guide will make more clear the evolving relationship
between ANS and the mid-levels who are noticably silent on com-priv. I
have the impression that for the remaining roughly 24 months of the
extended Merit agreement mid-levels which do NOT sign connectivity
agreements with ANS will remain connected to the backbone at no additional
cost to themselves but will be blocked from receiving ANS commercial
traffic. A point that is VERY important but rather less clear is whether
clients of the mid-levels which a "com" as their network address will be
able to send traffic which they state is in conformity with acceptable use
requiremenst accross the backbone. Does the hitchikers guide make this
clear?
>
>
Gordon,
Let me try to respond to several simple questions in your notes from
December 17th:
<> You understood *correctly* that "for the remaining ... months of the
extended MERIT agreement, mid-levels which do NOT sign connectivity
agreements with ANS will remain connected to the backbone at no additional
cost to themselves but will be blocked from receiving ANS commercial
traffic." One quibble is that it is possible that the mid-level and ANS
could agree on a variant of the connectivity agreement, but you've
captured an important point correctly.
<> You ask whether clients of the mid-levels which [have a] ".com" as
their network address will be able to send traffic which they state is in
conformity with acceptable use requirements across the backbone. The
answer is "yes" and "of course". The NSF AUP allows research/education
traffic to use the backbone without regard for whether the organization
sending/receiving the traffic is a university, kindergarten, government
research lab, corporate research lab, or consulting shop. Thus, whether
the suffix of the network's domain name is ".edu" or ".com" does not
matter.
<> In your second note, you ask about the difference between a Gateway
attachment and an Enterprise attachment. A Gateway attachment is one in
which we provide a connection between ANSnet and another network,
typically a mid-level. An Enterprise attachment is one in which we
provide a direct connection between ANSnet and a specific site, typically
a university or an industrial site. A Gateway can advertise to us network
numbers of its customers at many sites. An Enterprise can only advertise
to us its internal network numbers. The engineering of the two kinds of
attachments is essentially the same.
If you still have questions, feel free to give me a call.
Al (659) ------------------------------------------------
SUBJECT: Who Pays Whom How Much? A Murky Picture.]]
MESSAGE from =uunet!ans.net!weis 30-DEC-91 20:42
Received: from psi.com by nis.ans.net with SMTP id AA19829
(5.65c/IDA-1.4.4 for <com-privers@nis.ans.net>); Fri, 27 Dec 1991
02:51:26 -0500
Received: by psi.com (5.61/2.1-PSI/PSINet)
id AA04850; Fri, 27 Dec 91 02:52:34 -0500
Received: from relay1.UU.NET by psi.com (5.61/2.1-PSI/PSINet)
id AA03681; Thu, 26 Dec 91 23:43:46 -0500
Received: from uunet.uu.net (via LOCALHOST.UU.NET) by relay1.UU.NET with
SMTP
(5.61/UUNET-internet-primary) id AA08193; Thu, 26 Dec 91 23:42:09
-0500
Received: from tmn.UUCP by uunet.uu.net with UUCP/RMAIL
(queueing-rmail) id 234122.17812; Thu, 26 Dec 1991 23:41:22 EST
Received: by tmn.com (smail2.5)
id AA22713; 26 Dec 91 23:22:49 EST (Thu)
To: com-priv@psi.com
Subject: Who Pays Whom How Much? A Murky Picture.
Date: 26 Dec 91 23:22:49 EST (Thu)
Message-Id: <9112262322.AA22711@tmn.com>
From: cook@tmn.com (Gordon Cook)
<<MESSAGE from>> Gordon Cook 26-DEC-91 23:22
cook@tmn
Again what is unanswered in the Communications Week Dec 23rd article is
when the mid-level gets to sign the gateway or cooperative agreement. The
claim is made that all they are asked to sign is a coonectivity agreement
which apparently places them under no obligation except to receive
commercial traffic from ANS' customers. It even gives them something for
nothing -- participation in the infrastructure pool. I am having
difficulty believing that mid-levels can sign the connectivity agreement
(excuse typo above) and NEVER be asked to choose between the gateway or
cooperative agreements also outlined by ANS in its September documents. I
have asked privately about this and have so far received no clarification.
We read that a regional network that wants to send commercial traffic to
other regional networks over the ANS network has to pay ANS a fee. This
sounds to me like the gateway agreement. If I read that correctly such
agreements could be quite costly to the mid-levels. However I have yet to
see an example of exactly how much signing such an agreement would cost a
given mid-level.
The NSF has had to pay a more than 3 fold increase to ANS for the T- 3
backbone. It now appears that the mid-levels - if they are to keep their
commercial customers - are to also pay ANS for access to the backbone.Some
people have told me that they doubt` that the mid- levels that give up
their commercial customers can survive.
Can anyone offer some enlightenment on the conomics of all this?
--------------------------
Gordon,
You raise some interesting issues, several of which are very importan
to us at ANS. Let me try to answer the questions that ask for
clarificatio of ANS policies, and offer our hopes and intentions for the
questions that speculate about "the economics of all this"?
<> You first ask about whether a mid-level can "sign the connectivity
agreement and NEVER be asked to choose between the gateway or cooperative
agreements". The answer (simple in this case) is: yes, a mid-level can
certainly sign the connectivity agreement and never sign either the
gateway or cooperative agreement.
<> You state correctly that "a regional network that wants to send
commercial traffic to other regional networks over the ANS network to pay
ANS a fee". Then you speculate that "such agreements could be
quite costly to the mid-levels". Our intention is that the Gateway or
Cooperative agreement would be quite moderate in cost to the regional.
Assuming for a moment that we are talking about a regional with an
existing NSF-sponsored R&E service attachment and a situation in which the
regional's commercial traffic would initially be a small percentage of its
total traffic, we think that these costs would be quite modest. And, as
the regional's commercial traffic increases, we think these costs would
continue to be a small fraction of the regional's resources.
<> You mention that T3 costs NSF more than T1, then suggest that this
increased cost is being shifted to the regionals, then close with saying
that the regionals may not be able to 'survive' if they can't keep their
commercial customers. This combines three questions into one in a
way that makes a simple reply awkward. But let's try.
First, T3 does cost more than T1. That's not a surprise, and it's our
hunch that the regionals that are now about self-sufficient at the T1
level now would like to be self-sufficient at the T3 level in the not-
too- distant future. We think that support for mixed traffic across the
T3 network will serve the regionals well in this respect.
Second, you suggest that the regionals are being asked to pay for
this increased cost. It is important to stress that neither the Gateway
agreement nor the Cooperative agreement are needed for the continued use
by the regional of its NSF-sponsored service attachment for R&E use, i.e.,
for use in accordance with the NSF AUP. The only regionals being asked to
help pay for T3 are the regionals that choose to use their T3 attachment
for their commercial (i.e., non-NSF-AUP) traffic. These regionals have at
least three other alternatives: (a) they could get a separate T1 ANS CO+RE
attachment for their commercial traffic, (b) they could get a commercial
gateway attachment from someone else, or (c) they could decide to 'do
without' supporting commercial service. ANS doesn't take away
possibilities, but we try to create them.
Third, you speculate that regionals may not be able to 'survive' if
they have to 'give up their commercial customers'. The objective fact is
that the regionals (and their members) benefit from economies of scale.
And being able to support commercial traffic can only help in achieving
these economies of scale. For these reasons, we are trying to create
possibilities to make support by the regionals of their commercial
customers viable.
Our intent and hope in all this is to help the regionals move in the
direction of self-sufficiency at the T3 level using mixed traffic as a
means for the regionals to achieve the economies of scale needed.
If you still have questions, feel free to give me a call.
Al
_____ Note from Gordon Jan 3: I responded to Al saying that I thought it
would be helpful to take this public and test his hope that the mid-levels
who had reached t-1 self sufficiency were now interested in attaining T-3
self sufficiency. I observed that perhaps those mid-levels who agreed ...
or disagreed would say so on com-priv.
