From: dat@ebt.com (David Taffs)
Date: Wed, 11 May 94 17:05:25 PDT
To: lile@netcom.com
Subject: Re: So what do you think...
In-Reply-To: <199405112103.OAA18969@netcom.com>
Message-ID: <9405120004.AA10743@helpmann.ebt.com>
MIME-Version: 1.0
Content-Type: text/plain



I signed it. I have in the past received from Michael Ward a copy of a
paper describing congestion-based pricing models, which seem to work
really well for the kinds of things that everybody wants.

This letter seems to take such models into consideration, and seems to
be more cautionary than reactionary. Obviously, a metered model (like
phones) would be inappropriate for Internet, but is probably what
phone/cable companies would like to charge, even though it would
stifle usage.

This letter, while reacting against all usage models, seems to me
to leave the door open to more intelligent pricing models, such
as the congestion model (e-mail me for details). Thus, it seems
to address my concerns, and my fright at the idea of a conventional
usage-based model was sufficient to get me to agree to sign the
letter, in spite of the fact it doesn't call out congestion-based
models explicitly as an alternative. I specifically agree with all
the recommendations.

The congestion-based pricing model is essentially this (if I remember
it correctly): every packet includes how much it would be willing to
pay to be sent within a given time frame. The switch sends the packets
with the highest bids, but charging them each the amount of the
cheapest sent packet. Other packets either wait or get NACK'ed (I
forget what happens here). Note that zero is a fine amount to bid --
it just means you wait until the line frees up. Packets have an
incentive to actually bid the correct amount they would be willing to
pay, but don't get charged if they bid too high. People who care about
throughput pay enough to add enough capacity so there is always some
slack time. It really seems to me to work like a charm. I've got a
paper on this (with references to further papers) if anyone is
interested.


   Date: Wed, 11 May 1994 14:03:21 -0700
   From: lile@netcom.com (Lile Elam)

   about this letter? Would you sign it?

   -lile


   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   Lile Elam	    |  "Remember... No matter where you go, there you are."
   lile@netcom.com     |		
   Un*x Admin / Artist |			 Buckaroo Banzai
   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


   ---------- Forwarded message ----------
   Date: Wed, 11 May 1994 12:37:27 -0400 (EDT)
   From: Michael Ward <mike@Essential.ORG>
   To: niiregional-l@rain.org
   Subject: Letter to NSF re: Internet Pricing


   Distributed to TAP-INFO, a free Internet Distribution List
   (subscription requests to listserver@essential.org) 

   TAXPAYER ASSETS PROJECT - INFORMATION POLICY NOTE
   May 7, 1994

   -    Request for signatures for a letter to NSF opposing metered
	pricing of Internet usage

   -    Please repost this request freely

   The letter will be sent to Steve Wolff, the Director of
   Networking and Communications for NSF.  The purpose of the letter
   is to express a number of user concerns about the future of
   Internet pricing.  NSF recently announced that is awarding five
   ...