[1913] in Commercialization & Privatization of the Internet
Re: Understanding Combits
daemon@ATHENA.MIT.EDU (Lars Poulsen)
Tue Jan 7 20:24:50 1992
Date: Tue, 7 Jan 92 17:14:33 PST
From: lars@casper.CMC.COM (Lars Poulsen)
To: sean@dsl.pitt.edu
Cc: com-priv@psi.com, jgong@us.oracle.com
I was suggesting that the CO/RE distinction is more trouble than it is
worth, and maybe the midlevels and backbone should just go commercial.
If NSF still wanted to underwrite the RE part of the network, it could
be done with a subsidy to qualifying RE subscriber institutions, who
would be free to shop around for service providers.
JG> From: "John Gong" <jgong@us.oracle.com>
JG> Hear, hear. Put the money into providing more accessible services and
JG> information. Then customers will decide whether they're willing to pay
JG> for it (they will) and how fast/how much they want (pipe size-oriented fee
JG> schedule).
I was actually not suggesting anything as radical as that. My proposal
extends only to the transport network. Who provides information services
and how they get paid is a separate, but also interesting question.
SML> From: sean@dsl.pitt.edu (sean mclinden)
SML>What about *essential* services. What is available on the Internet, now, is
SML>not just a business tool, it is an educational tool of tremendous value. It
SML>is all the technical journals, plus Consumer Reports, plus a lot of other
SML>things rolled into one. To talk of it as simply an e-mail/USENET news reader
SML>is, I think to choose a very poor metaphor.
SML>The computer (where, as Sun says "the network *is* the computer), is the
SML>the most valuable educational tool yet produced. It allows you to educate
SML>yourself at whatever pace you can sustain (or, in your world, whatever
SML>pace you can afford). Would you want your children's education to be billed
SML>by the minute (or by the byte) independent of what they actually used?
The commercialization of the longhaul network has only a peripheral
relationship to educational end-user costs. It is entirely possible
that we might end up with a system where the per-packet and per-byte
cost is negligible, allowing libraries to provide free walk-up terminals
for their users.
I am very concerned that there be affordable access for personal
educational use for ordinary people. I am just not convinced that
the most cost-effective way to achieve this is by a massive funnelling
of subsidies from big government (NSF) to big business (IBM/MCI).
If all interstate IP carriers were required to join CIX, or something
similar, and each of them was free to develop its own rate structure,
we would see a few years of intense competition between MCI, ATT, SPRint,
UUNET, PSI, CERFnet and the other FARnet members. Whether the winner would
be a wireline carrier, the FARnet regionals, or the nationwide IP
specialists is not at all clear to me. On the other hand, it does seem
pretty clear that if a dominant carrier is guranteed to get paid for all
the government sponsored RE traffic, that carrier has a unique competitive
advantage.
SML>And what about "dialing in." That may be sufficient for you, but people
SML>(and I mean normal citizens) are putting whole networks in their homes.
SML>Although groups such as the CODA group at CMU are working on disconnected
SML>mode filesystem, a great many Internet applications require connected
SML>mode to maintain reliability. Perhaps, in the future, disconnected
SML>operations supported on switched services will be the way to share a limited
SML>bandwidth, equitably. But that isn't what's running on the campuses of
SML>CMU, Berkeley, Stanford, MIT, or scores of other places.
I expect ISDN B-channels to become the standard residential attachment.
We should focus our efforts on making sure that those are reasonably
priced (i.e. at roughly the price of today's telephone connections)
for local calls.
I speak from, but certainly not for, CMC (a Rockwell International Company).
/ Lars Poulsen, SMTS Software Engineer
CMC Rockwell lars@CMC.COM