[10746] in Commercialization & Privatization of the Internet

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

Re: Settlements

daemon@ATHENA.MIT.EDU (Dan Lynch)
Mon Mar 7 18:32:01 1994

Date: 7 Mar 1994 12:12:17 -0800
From: "Dan Lynch" <dlynch@interop.com>
To: com-priv@psi.com, "Craig Partridge" <craig@aland.bbn.com>
Cc: dlynch@interop.com

        Reply to:   RE>>Settlements
Gee whizz, Craig,  I hear the old Tenex Pie Slice Scheduler coming back...  It
might work.  I loved its simplicity.  (For those of you who are blissfully
ignorant of that old technology, it was a timesharing scheduling system that
basically sold percentages of the clock of the CPU for a fixed price to any set
of users.)  The predictability of costs for the buyers and of revenues for the
providers was wonderful for both.  The only problem was for the provider to
keep the "response time" acceptable to the buyers.  (Buy more machines was the
response when the loads got too high.  It is the equivalent of "buy more
backbone" in this case.)

So, I like the idea.  I think we have come across a way to do "backbone
payments" with out doing it "by the byte".  Lots more to be discussed, and this
is probably not a good list to carry on this discussion.

Thanks,
Dan

--------------------------------------
Date: 3/7/94 7:34 AM
To: Dan Lynch
>From: Craig Partridge

> What about the following idea:  What is missing here is a true "IXC" for IP
> traffic.  A "carrier" whose only purpose was to take all "local" traffic and
> forward it to the appropriate "local" destination?  Forgetting who might own
> such an entity and forgetting the perhaps idiotic "Star" topology that that
> business model might look like, isn't that the place where settlements would
> need to happen?

Dan:

Let me play devil's advocate back.

Suppose the IXC simply billed each connected provider a flat amount
per month based on the size of the provider's pipe (or pipes) into the
long-haul network.  So, for instance, if PSI had T1 connections to the IXC
in LA and DC, PSI would pay for two T1 connections.

The IXC could make guarantees about loss rates and service outages, so
you'd know what you were buying.

The result:

    * a nice simple, predictable revenue stream for the IXC

    * a simple bill for the local carriers

    * possibly lower costs (less accounting work at both ends)

What's wrong with this picture that we have to inject settlements into it?

Craig

------------------ RFC822 Header Follows ------------------
Received: by interop.com with SMTP;7 Mar 1994 07:34:22 -0800
Received: from uu2.psi.com by polaris.interop.com (4.1/SMI-4.1)
	id AA20935; Mon, 7 Mar 94 07:34:45 PST
Received: from port13.sunnyvale.ca.pub-ip.psi.net by uu2.psi.com
(5.65b/4.0.071791-PSI/PSINet) via SMTP;
	id AA29052 for dlynch@interop.com; Mon, 7 Mar 94 10:30:23 -0500
Received: from localhost by aland.bbn.com (8.6.4/3.1.090690-BBN)
	id HAA07931; Mon, 7 Mar 1994 07:29:16 -0800
Message-Id: <199403071529.HAA07931@aland.bbn.com>
To: com-priv@psi.com
Cc: dlynch@interop.com
Subject: re: Settlements
In-Reply-To: Your message of Mon, 07 Mar 94 06:14:10 -0800.
             <199403071414.GAA07866@aland.bbn.com> 
>From: Craig Partridge <craig@aland.bbn.com>
Date: Mon, 07 Mar 94 07:29:15 -0800
Sender: craig@aland.bbn.com





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