[1837] in Commercialization & Privatization of the Internet

home help back first fref pref prev next nref lref last post

Who Pays Whom How Much?

daemon@ATHENA.MIT.EDU (Jordan Becker)
Thu Jan 2 18:35:54 1992

Date: Thu, 2 Jan 92 18:26:45 EST
From: Jordan Becker <becker@ans.net>
To: cook@tmn.com
Cc: com-priv@psi.com

> <<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 connectivity agreement 
>  which apparently places them under no obligation except to receive 
>  commercial traffic from ANS' customers.

Not quite.  When a peer network signs the connectivity agreement, they
agree to support the exchange of traffic bidirectionally between R&E
classified networks hosted by the peer network, and commercial networks which
are reachable via ANS.  Networks are classified by the acceptable use policy
that they choose to conform to.

> It even gives them something for nothing -- participation in the
> infrastructure pool.

Not really, they have to carry the commercial traffic within their
infrastructure, which costs them something.  This is the reason for 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.

They might be asked to sign the gateway or cooperative agreements if they
wanted to classify networks using the ANSNET which they host as commercial
rather than R&E.  Charges for commercial traffic would be assessed.  This is
what funds the infrastructure pool.
  
>  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.

You are correct.

>  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.

Please see the ANS plan for Commercialization which may be FTP'ed from
ftp.ans.net.  There are examples contained in these documents.

>  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.

Again, a network is classified as commercial if it chooses not to adhere to
the NSF (or ANS R&E) acceptable usage policy.  Many commercial institutions
are quite comfortable with the NSF AUP, and are unaffected by any new
agreements proposed by ANS.

If they are not comfortable with the NSF AUP, then we offer them an
alternative for which they may elect to pay a few extra dollars on behalf of
their commercial subscribers so that federal funds are not used to subsidize
their otherwise unacceptable use.  All of these 'extra dollars' go into the
infrastructure pool which benefits other midlevels that participate in the
connectivity agreement.

Even if ANS was willing to put the 'extra dollars' in the infrastructure pool
on behalf of the commercially classified networks that can not pay, this would
still not give us the right to transport this traffic into midlevels that can
not support it for legal or other reasons.  This is why the connectivity
agreement is desired and, midlevel consent is required as a minimum.

This model was developed in consultation with many members of the Internet
community and has been iterated based upon the constructive feedback we have
received.  The goal has always been to support the introduction of commercial
services (where being commercial is self-declared by the customer) in a way
that allows co-existence with government sponsored services dedicated to R&E
use.  Any constructive suggestions on further iteration of this model are
always welcome.


	Jordan

home help back first fref pref prev next nref lref last post