[2121] in Commercialization & Privatization of the Internet
Re: Trickle Down? How About Bubble Up?
daemon@ATHENA.MIT.EDU (Edward Vielmetti)
Mon Jan 27 19:15:03 1992
To: stef@nma.com
Cc: com-priv@uu.psi.com
Date: Mon, 27 Jan 92 19:03:57 -0500
From: Edward Vielmetti <emv@msen.com>
> Well guys -- It sure looks clear now that IP/PPP is still in the
> experimental stage, and that the claims that it is better than
> somethign else are still speculation.
Maybe experimentation for you, but here in Ann Arbor it's pretty
much plug in and go. A V.32bis modem gives equivalent performance
to a 19.2 kb/sec LADS circuit on real world data. With either of
the Livingston Portmaster terminal server or the Morningstar PPP/SLIP
code the connection works more or less right out of the box, and
the support organizations behind these products are quite good.
The real hassle to be overcome is not the serial line encapsulation -
PPP is by any argument better, but SLIP can be quite adequate - it's
the necessary configuration and management code on each side to bring the line
up when it's needed and tear it down when the link is idle.
I fully expect that this technology will bubble up to higher speed
switched services as time goes on.
> his is wonderful, and I can hardly wait to see the results.
The results should be troubling for traditional slow-moving
network organizations that expect a big downpayment to get users
connected. The capital costs of a network setup are going way
down, to the point where it's not out of the question to have
people at home be on the internet. Traditionally that was only
possible within the confines of a big company or a big university.
(Or in some very select places on either coast.)
> In the meantime, back to SENDMAIL/UUCP FIDO and MMDF/PHONENET
These are each appropriate technologies for certain problems.
Intermittent connections still probably want a store and forward
scheme for mail. Hopefully you can end up with a hybrid scheme
where you have part time IP connections sharing the same modem
infrastructure with store and forward UUCP stuff, bringing the
marginal cost of adding another system down even further.
The real issue in all of this is policy and regulation, *not*
technology. Many current mid-level connection policies actively
discourage development of these sorts of services, either by
pricing the lowest end connection out of the market or by simply
refusing to attach people who want to build nets like this.
(CIX members seem to be better at this though even there there
are known problems with some early efforts.)
--Ed
