Xref: utzoo news.admin:10912 news.misc:5615 comp.protocols.tcp-ip:13559
Path: utzoo!attcan!uunet!zaphod.mps.ohio-state.edu!usc!bbn.com!cosell
From: cosell@bbn.com (Bernie Cosell)
Newsgroups: news.admin,news.misc,comp.protocols.tcp-ip
Subject: Re: Internet/NSFNet proposal to be run by IBM -- call to action!
Message-ID: <60411@bbn.BBN.COM>
Date: 29 Oct 90 14:24:14 GMT
References: <1990Oct28.220432.521@ddsw1.MCS.COM>
Sender: news@bbn.com
Followup-To: news.admin
Lines: 161

karl@ddsw1.MCS.COM (Karl Denninger) writes:

}This is a call to action by all interested parties.

}There is wind of a proposal stirring in Washington that would place the 
}NSFNet backbone, and eventually the entire government-run part of the 
}Internet, into the hands of IBM.

..

}Anyone want to bet how long the Internet remains accessible to non-IBM
}people?  

I'll take that bet --- I suspect this would never happen.

]Or whether the Internet ends up another Prodidy, with active
}censorship?

This is a possibility, but 'censorship' is a bit strong here.  Fact is
that we're using the network's comm resources for purposes quite a bit
beyond the intent of the folks who chartered the thing.  This is,
sooner or later, going to come to a head: either the screws will be
tightened, restricting the network to its narrowly-defined actual
purposes [the case that you fear], or else some set of folk will manage
to get the 'powers-that-be' [actually it is the 'powers-that-pay'] to
acknowlegd the utility of the network in its broader uses.

I, actually, think that the latter is fairly unlikely: as the gov't
financial world gets more constrained, it will prove harder and harder
to defend the somewhat amorphous 'general good' served by running the
communications service, especially in the face of _concrete_ research
projects having to be discontinued for lack of funds.

Whether you'll have to buy an IBM system to hook into it, 
}since they might decide that TCP/IP is no longer any good and now it's 
}time to go to SNA or worse?

This is spectacularly unlikely.  If anything, TCP/IP will be abandoned
in favor of ISO stacks, but even that is not going to happen any time
soon.

}Folks, if you love the Internet, and want to see it expand and grow, we need
}to insure that a few things happen:

}1) 	The "acceptable use" policy on the NSFNet needs to be scrapped.
}	Sure, this will bring problems.  But it will also mean that
}	commercial companies can tie in, pay their fair share, and make sure
}	that the network capacity has the funding to continue to grow.

What would you replace the policy with?  I think that the NSF charter
constrains their budget to be used as grants to further research.  I'd
bet that they can divert funds to running a communications service only
as long as such a service more-or-less directly supports their research
mandate.  [I know that we had such a problem with ARPANET funding, way
back when: when it became obvious that the ARPANET, per se, had become
a communications service and not a subject of research in and of itself
[indeed, if we did any real messing-around with the ARPANET we got
near-instantaneous complaints from all over the world! :-)], ARPA got
unhappy about funding it for precisely those two reasons: (1) it was
beyond ARPA's mandate, and so would not withstand Congressional
scrutiny, and (2) it preempted funds from other (real) research
projects.  The result was that the ARPANET operation was transferred to
DCA,a nd eventually resulted in the ARPANET being decommitted entirely
[the argument being that that sort of general comm facility was not an
appropriate activity for the gov't to be running]


}2)	The backbone needs to be run as a regulated commodity.  Perhaps even
}	run by the Government, strange as that may seem.  The goal of
}	universal connectivity is not that far off right now, but there are
}	companies and special interests who would like to see that never
}	happen.  We MUST insure that it does.

I think you'll have a VERY hard time making a really rational case for
this.  There are commercial alternatives to the gov't-subsidized
internet, and one might argue that there might well be more of 'em, and
they'd be more economical, if the gov't weren't gigantically perturbing
the market by making so much capacity available 'for free' [that is,
'paid for by someone else'].  The easiest way to achieve your goal [of
universial connectivity] is to be exploring ways to PAY for the sort of
communications services you envision, rather than just thumping the
drum that the gov't should provide it.


}4)	Finally, a freely accessible information exchange medium may be the
}	second-best guarantee of freedom in this country (the first being
}	the ability of the people to depose a despotic government).  By
}	keeping the passing of information from coast to coast available, 
}	fast and cheap, we keep the people informed.

Granted.  The only debate is the perennial one: WHO SHOULD PAY.  Why not
argue for a self-supporting communications service, instead of a
taxpayer-supported one.  The problem with the gov't-run solution is the
obvious ones that crop up whenever you let the gov't run stuff:
    1) if they pay for it, they'll expect to have some say about what it is
       used for.
    2) if they pay for it, they'll make blanket rules about how it is glued
       together
The problem with (2) is that (via the taxes we pay) we're all going to have
to subsidize this communications facility even if it doesn't serve our
purposes.  As for (1), one need only look at the 55mph stuff and the DFWF
extortion programs to see how overall nasty it can be to suckle at the
gov't's teat.


}	1)	A tax on access devices for the network may be the best
}		way to fund it.  I'm not sure about this, but it seems as
}		though a "user fee" is one of the better ways to pay for the
}		connectivity that we all enjoy and want to see furthered.

If this is adequate to PAY for the network [I cannot imagine that it is ---
do you have any clue how expensive the operating costs for something like
this is?], why not simply have it be privately operated?  Then the gov'ts
opinion on what the net does, and what people send over it, becomes
irrelevant.

}	2)	General subsidy isn't a bad idea either, but it's not ideal.
}		Selling it to the general public will be difficult,
}		especially with the things that hit the press now and again
}		about X-rated GIF sites and the like.

Bingo.

}	3)	Keep control in the hands of the many, or in the hands of a
}		non-profit corporation funded EXCLUSIVELY to run this beast.
}		Giving it to IBM or another pseudo-government company is as
}		good as letting the fox loose in the henhouse -- the
}		potential for abuse and profiteering is just too great to
}		ignore.

Perhaps.  Going back to your original 'the sky is falling' cry, you
only say "... into the hands of IBM".  What does that mean?  In the
past, in the various networks that have contracted out to be 'run'
[including the long-standing contract from ARPA, and later DCA, to BBN
to operate the ARPANET and the MILNET, and one from NSF to operate the
old CSNET] have been quite clear as to the responsibilites and
prerogatives of the operations-entity.  I cannot believe that NSF
either would want, or would be legally allowed, to turn over the policy
and planning of the NFSnet to IBM: the purpose of the network would
still be as it was [to serve NSF's research interests] and only the NSF
can evaluate and balance those purposes.

It feels to me that you're doing two things here, neither of which
serves us well: first is blowing the "danger" all out of proportion.
Second, I think you're vastly underestimating just how much it costs to
run one of these things. [and so you're being a bit cavalier at just
how easy it would be to find enough money to make this 'non profit
corp' to actually be able to pay its bills].

Instead of pushing for MORE gov't involvement in usenet, what is wrong
with pushing for LESS: how about we eschew those 'free' govt-supplied
links [which we really, probably, oughtn't have been using for rec.pets
in the first place] and instead push to have more "pay for play"
links?  If we make our own communications arrangments [which could
include using Telenet, or even, as you suggest, starting some kind of
non-profit corp to provide some kind of backbone facilities], in
addition to getting the connectivity you want, we have two BIG other
advantages: (1) we don't have to keep groveling before Congress for
funds, and (2) we don't have to duck reporters.

  /Bernie\