On 3 Feb 99, at 10:29, Ivan Pope wrote:
> >I've been associated with the Internet for a VERY VERY long time, and it
> >has -never- made sense to me that it has never had a proper cost recovery
> >model that involved paying for what you use (I lay this problem on NSF's
> >doorstep; the situation was arguably different for DCA and ARPA, but
> >there was never any excuse, IMO, for NSF not to put in proper
> >accountability/accounting). What's interesting is that it'd be
> >relatively easy to do: every router and gateway already keeps enough
> >statistics to be able to generate bills for the next person 'downstream'.
> >Scary/intriguing thoughts, but it ain't gonna happen, more's-the-pity,
> >and so we're still left with the mess...
>
> I think that if there had been a 'pay for all use at the individual point'
> system, there would be no Internet.
It is water long over the dam, but I think you're wrong. If, when NSF
started up CSNET [and later set up the NSFnet backbones and the
regionals], they had set up a billback scheme, so that schools got billed
according to the traffic they generated, I don't think the Internet would
have died at all, any more than that schools don't provide phones because
the phone company bills them for usage, or that they don't subscribe to
journals because they have to pay to receive them. I think it would have
worked out just fine, but certainly been different. While the 'billing
machinery' was small and a the community relatively closed [since in the
end NSF was picking up most of the bills anyway via grants] there would
have been ample time to sort out various policy matters [how to pay for
mailing lists... how to pay for FTP archives, and then later how to pay
for IRC, how to pay for usenet, how to pay for the web, etc].
I don't think the per-transaction cost for normal things would have
worked out to being very large [how much can it cost to send a 1K email
message on a T3?] and would have only discouraged the applications that
were REAL bandwidth hogs and were arguably useless [like the clowns at
MIT who walked around with video cams on their heads all day, or the real-
time-video of some student's pet iguana]. Even then, schools could get
grants to research whateveritwas and just pay for the bandwidth [just as
they have to actually budget for, and pay for, every OTHER resource they
consume, from pencils to photocopies to trips to conferences] for
research they fund. The MIT media lab had to pay for *EVERY*OTHER*
resource it consumed, -except- the network communicatiosn bandwith its
applications ate; why should that one addition resource/expense have been
a show stopper?
Places like Amazon and buy.com and friends would trivially pay the minor
transaction costs to run their web sites [although they might thing twice
about the fancy graphics]. Just think how it would've been if the folk
that designed X-windows had been doing so with a mindset of pay-for-the-
bandwidth-it-uses rather than "everyone using X will be on a 100meg LAN".
Places like NPR [and other 'real audio' providers like ABC, CNBC, etc]
routinely *expect* to pay for all of their other communications costs
[satellite links, 800-dialins, etc], why would having to pay for their
internet services be a make-or-break situation?
My guess is that personal accounts would end up being like today's cell-
phone accounts are: there woudl be a LOT of different plans, but instead
of having just "connect hours" [and sometimes "web page storage"] bundled
in, they'd also have some "base bandwidth". For your $19.95 a month, you
get 5 megs/month of data transfer or some such.
I think it all would have been doable [indeed, just barely, conceivably
_still_ could be doable] and I don't think it would've killed the
internet at all. The key things that make the internet work are *NOT*
[IMO, of course] the huge bandwidth hogs.
> ... There would of course be proprietory
> networks for business etc and people would still be trying to get us to
> subscribe to various services and no doubt we would.
I think you've got this wrong. There already *WERE* proprietary networks
pre-CIX/pre-regionals. Prodigy, CIS, etc. They were *dieing* to
interconnect with the "real net" and I just don't believe that a pay-for-
usage model would have deterred them. THey might have some kind of two-
tier setup [which, in fact, AOL had for quite a while: AOL-local email
was free, but you paid for internet email [but they got it wrong, of
course, since you paid to *receive* internet email]].
> Of course, every business does pay for its traffic. One of the reasons the
> Internet has worked so well is because its made us all think very hard
> about how to pay for the services that we know the Internet should provide.
Foo!!!! One of the reasons the internet has worked so well is
*just*the*opposite*. Folks approached it with a "bandwidth is free, how
can I make a buck" attitude or at the least with a "bandwidth is free,
what neat/fun/useless thing can I *do* with that bandwidth" [and so you
get InternetPhone and the Fax handlers and folks selling software and
discovering that they can save $10 on each item in
handling/admin/shipping costs because sending their stuff out over the
internet "is free", not to mention iguanacam and friends].
> But, to say that we should have had 'pay for what you use' from the start
> misses the point that there would be no Internet if that had been the case.
I think we're doomed to disagree on this, but you might expand on why you
think that's the case rather than just saying it ex cathedra. I actually
think the Internet would have progressed *exactly* as it did, only
perhaps a bit more slowly, a bit more deliberately, and with a bit more
attention to utility/bandwidthconsumed.
The gov't sites would've been on the net anyway, and their internet costs
would just be budgeted just like every other cost they incur now
[printing, phones, personnel, etc]. the schools certainly would've been
on, as I outlined above. Maybe we wouldn't have MUDs, maybe not IRC [or
perhaps it woudl've been different], but they'd all be there.
The schools that leaned on NSF to start CSNET would certainly not have
backed off if they were just given a bill for what they used --- schools
already had to deal with that for EVERY other resource they provide for
their students and staff, from photocopiers to journal subscriptions to
conferences to research projets. BUT: what it woudl have done would have
almost certainly made schools be -responsible- for their use, and cut
down on the really crazy "let's eat more bandwidth because it is fun and
free" hacks they pursued [where iguanacam is probably my favorite
exemplar for that kind of thing, and just think that there might not have
been as much of a "september effect"]
Businesses were *dieing* to get connected and they would surely have
connected right up as soon as it became legal. It would have been a bit
different of course: more services like the handful of "subscription"
services we have now and fewer of the "take all you want for free"
ftp/document sites, but even for that many companies would realize that
their handling/shipping costs would be so much lower via the Internet
they'd probably eat the costs anyway. Certainly the 'retail' places
would [the buy.coms, amazons, cdnows, etc].
The only potential question for me is "civilian" presence. Would there
be the flood of everyone and their aunt to the Internet? I think the
answer is yes: I think that "ISP accounts" would come bundled with more
"free" bandwidth than almost any normal user would use and I also don't
think [because I have a LOT of faith in the resourcefulness of the
business world] that they'd be put off because everything *on* the
internet would be so expensive. Maybe they'd have to pay a small fee to
"browse the Louvre", but what's the problem with that?
And think about some other things: if sysops got *billed* for their net
usage, do you think there'd be many open/anonymous servers misconfigured
out there? And so in addition to [IMO] doing a lot to cut down on spam,
it'd also [again, IMO] have put a BIG crimp in the hacker-attack
problems.
/Bernie\
--
Bernie Cosell Fantasy Farm Fibers
mailto:[EMAIL PROTECTED] Pearisburg, VA
--> Too many people, too few sheep <--
