Re: The FAQ system approaches obsolescence. What do we do now?

---------

Ross Smith (alien@meanmach.actrix.gen.nz)
Tue, 13 Dec 94 08:00:57 GMT+12


Eric S. Raymond wrote:

> > Please tell that to all the ISP providers in New Zealand. Everyone here is
> > charged on a volume basis for IP packets travelling over our international
> > link. I know of only one service provider who is absorbing this cost and not
> > passing it on to the user and that is a City Council run CityNet scheme.
>
> Interesting...someone finally raises what really looks like a real exception.
> Causes me to suspect rather strongly that there is a monopolist somewhere in
> the supply chain. Is NZ's telco a state-protected monopoly?

No, but NZ's telcos aren't involved, except on a local basis. The
international net link *is* a de facto monopoly, i.e. nobody else seems
to be interested in competing (apart from companies like IBM that can
charge more because they have a captive market). It's run at least partly
(I don't know all the details) by Waikato University, and isn't making a
serious profit (maybe none, I'm not sure). International links cost real
money, and that gets passed on.

> > Have phone companies gone to allowing unlimited toll calls for a flat
> > monthly fee yet?
>
> No, that's why I specified *local* phone service.

>From what I hear, more and more local phone providers (in the US and
elsewhere) are moving to charge-by-time for local calls. NZ Telecom
wants to, and probably will once the public has got used to the idea.

> > Your arguements rely on a) their being effectively unlimited networking
> >resources available to satisfy the customers demands (to prevent users becoming
> >dis-satisfied with the flat rate fee they are being charged.) & b) a large user
> >base to spread the load over and make the average cost drop to managable
> >levels. I don't see this happening in a hurry inside New Zealand.
>
> On the face of it, you make a strong case. I almost believed your implied
> argument that the non-viability of charge-by-byte is contingent on a) and b).
> Then it occurred to me that this theory fails to retrodict -- it doesn't
> explain why charge-by-byte wasn't viable in the U.S. back when the U.S.'s
> Internet infrastructure was comparable to NZ's.
>
> Can we develop an explanation for this?

Simple. There never was a time when the US's Internet infrastructure was
comparable to NZ's, because, as someone else put it, you got handed a nice
fat backbone on a plate by ARPA and the NSF. The point I've been trying to
make is that, as the backbone moves from non-commercial and non-profit to
commercial and out-to-gouge-all-they-can-make, charging by the byte will
become an increasingly attractive option to the suppliers, and eventually,
the only option.

Think for a moment in terms of the distribution of usage. In any system
with a natural minimum (zero) but no natural maximum, the distribution
is going to have a long tail (i.e. a few really-high-volume users and a
lot of medium-volume ones). That means the mean usage (on which charging
by time would have to be based) is going to be greater than the median.
It follows that charge-by-time has to be more expensive, on average, for
most users than charge-by-byte. (If you don't believe me, work it out
for yourself. Assume a realistic distribution of usage, and two providers,
both paying the same amount for their bandwidth and making the same
profit. One charges by the byte, the other by the second. Most of the
consumers, you'll find, will prefer the first.)

That's why charge-by-time was so unpopular when it was suggested here.

> > > Free markets. Ya gotta love 'em.

Not a popular sentiment in New Zealand these days...

... Ross Smith (Wellington, New Zealand) <alien@meanmach.actrix.gen.nz> ...
Keeper of the FAQ for rec.aviation.military
"I inspected the man closely -- he was the nearest thing I'd seen to a
human being, without actually being one." (Ned Seagoon)



[ Usenet Hypertext FAQ Archive | Search Mail Archive | Authors | Usenet ]
[ 1993 | 1994 | 1995 | 1996 | 1997 ]

---------

faq-admin@landfield.com

© Copyright The Landfield Group, 1997
All rights reserved