[2453] in Commercialization & Privatization of the Internet
Re: Billing Philosophy
daemon@ATHENA.MIT.EDU (HORN%athena@leia.polaroid.com)
Sat Feb 29 11:51:58 1992
Date: Sat, 29 Feb 1992 11:53 EST
From: HORN%athena@leia.polaroid.com
To: com-priv@psi.com
This billing discussion and some previous discussions are the tip of the
iceberg of utility economics. It is time for all of us to really study the
past and present of telecommunications, electric, and railroad regulation,
deregulation and pricing. I include railroad because it is both relevant and
the only one with really substantial history of alternative technologies and
commercial competition. The discussions here are the equivalent of the early
``send packets with checksums and resend the failures'' discusions that
preceded TCP/IP (e.g. pre-NCP).
In that spirit of acknowledged gross over-generalization, I offer these
comments.
If your goal is the most cost efficient use of resources to deliver service ---
(Note the IF. This is commonly the case with commercial products but many
reasonable people disagree with this goal in practice.) --- you want your
pricing to approximate that of ``pure competition'' which in turn approximates
the actual cost. This leads to the following definition of usage based
pricing:
Usage pricing reflects the actual cost of providing that service
So the metric of usage should reflect actual cost. The telephone industry
evolved under a political process with a very different technological base.
The typical telephone usage is about 99% idle with bursts of 56Kb/s continuous
use. Between the bursty usage and the political pressures, usage came to be
defined as connect time regardless of the actual cost factors.
My experience with private LANs and WANs, and the experience related to me by
Email, indicate that the major cost factors in providing Internet service are:
1) Network operations labor, (driven by service level and capacity, not
connect time or packets)
2) Capacity provision, e.g. lines and routers.
Therefore the present system of using capacity as the basis for billing is a
rational approximation to the actual cost. It should be no surprise that the
dominant trend for Internet service providers is capacity based pricing. I
would expect the next trend to be a distinction by level of service rather than
by connect time or packets, since that seems to be the next largest factor in
operating costs.
I personally would describe today's situation as more approximating ``price
fixe' '' and not all you can eat. When I purchase Internet services I get a
small selection of fixed services and capacities chosen by the provider. This
is a perfectly valid and successful financial model in numerous industries, so
why not the Internet?
Rob Horn horn%hydra@polaroid.com
