[11455] in Commercialization & Privatization of the Internet
No subject found in mail header
daemon@ATHENA.MIT.EDU (Anonymous)
Sun Apr 3 05:21:12 1994
Date: Sun, 3 Apr 94 00:41:29 -0500
From: Anonymous <nowhere@bsu-cs.bsu.edu>
To: com-priv@psi.com
karl@mcs.com (Karl Denninger) writes:
> What could happen with, say, $200,000 in additional annual
>income for the CIX? I can think of a couple of things right off the bat:
>1) New points of service. We could have a CIX-West, CIX-East,
> CIX-Midwest and CIX-South. All connected with T1 lines. NOW the
Obviously CIX could do things like this with $200,000 extra in annual
income. So could the ISPs whose money that is you are planning to spend
(and they probably would prefer to decide voluntarily how to spend their
own money rather than let you decide for them). It could be used by other
companies that might wish to provide interconnection points competitively
to the CIX. If they choose to they can buy into this vision and this
provider, CIX. Or perhaps they might wish to form CIX2 with their money.
Or connection points in different places from where you wish to place them,
perhaps a provider wishes to try to get one near where they are and wants
to put his money to that use. Perhaps you can persuade people to buy into
that vision, but if so ask them to do so voluntarily and with an
understanding and agreement with the value they are getting for their
money, not simply because they should pay in so they can get "guaranteed"
routing (without a good argument for why they should fear for their
routing).
I actually wouldn't be surprised to see some competitors to CIX in terms
of connection points, and other private backbones begining to arise.
People always seem to be able to think of ways to spend other peoples
money for them, and seem to feel they know better how to spend it than the
people they wish to take it from.
> More service. As membership grows, those T1s could become T2s and
> even T3s. Imagine a 45mbps AUP-free <cooperative> backbone, with a
As will be the case in general with competition driving the prices down,
and with other providers whose infrastructure will be built on selling to
willing customers (including the CIX if they start only charging direct
connects). You don't need to try to make CIX into a monopoly to get these
things, if anything competition will be more likely to make them come into
being, there will be creative things people will do to create alternatives.
I've heard stories from some people about ways they have discovered to get
T1 long haul telecomm lines (and presumably T3 as well though I didn't ask
at the time) for an incredibly tiny fraction of the price people usually
pay for them because of all the unused fiber out there, if you know where
to buy it from and how to get past the usual pricing structure. I can't
give details on this since I don't know them at the moment, and I haven't
dealt with this stuff directly myself. However you will start seeing the
results later this year probably.
>Those ISPs who have made their business selling "resale" connections
>(primarily
> ANS) are in big trouble in this situation -- is it any wonder they
> would oppose this? All of a sudden you have available for $10k a
> year what ANS would like to charge you $70,000 for!
Just as it is no wonder that existing local/regional ISPs with money to
throw around might oppose the coming hoard of mom and pop ISPs
bootstrapping from PCs and low budgets and threatening to come in and add
competition.
>What are you going to do if the following happens?
The case presented here was dependant on several assumptions which were
unlikely, such as companies forgoing selling to resellers, which seems
questionable. For instance Sprint is basically in the long distance
business, why wouldn't it try to make money selling to locals? If there is
a market that CIX or other companies abandon for resale to ISPs, are you
sure that no one else will step in to fill their place, given the potential
money and the growth that everyone seems to see in this area? If there is
money to be made, why wouldn't they? The links can be bought, and there are
enough companies with deep enough pockets that it would be done. At the
very worst case it would start out with T1s among the small ISPs
themselves. It really appears to be a straw man that doesn't make sense,
just a scare scenario. And we might get hit by a meteorite also. Perhaps
there is more chance of what you describe happening, but I'd like to see
rational arguments as to why it would happen and why no one else would step
in. You yourself mentioned that if there were a threat of this you would
step in and set up a CIX in the midwest. CIX isn't and wouldn't be the only
solution, or there is something seriously wrong with existing companies,
or with investors and entrepreneurs these days. We don't need to latch
desperately onto the first thing that some people feel is the way to go
when this industry is really just taking off now and probably isn't in
danger of heading back to the way things used to me. Especially not without
rational discussion of the alternatives.
Its premature to assume there won't be other companies and providers
getting into the interconnection point and backbone network area, simply
because it hasn't been done yet. It especially takes a while for companies
to decide to risk getting into an area the government has been involved in
before (and is still unclear over its role in) and where they fear the
government treading into again and competing with them or over-regulating
them, etc.
>Instead, you have to negotiate 30 agreements with 30 other
> people, and pay all of them.
Why are you assuming that this will be the case (even if it perhaps was at
some point in the past, or threatened to be at least)? Why couldn't it be
the case that you have a very low price place to provide cooperative
routing agreements for a small filing fee for signing the contract? If all
the providers benefit from a central clearinghouse, then it will happen.
Perhaps another is in the works now, I don't know. If places are paying
$10k/yr and deciding its in their self interest to have cooperative routing
agreements with the CIX, will they suddently become irrational if they can
get that same thing for $50 once? Isn't it possible that if the existing
ISPs decide to adopt a rational policy of having agreeing to route traffice
between all of the pipes connecting to the ISP, that this policy becoming
the defacto standard would let the industry work? Think about it, the
reason you pay to connect to a backbone would be that it routes traffic
among all the networks connected to it. If you connect to a regional/local
ISP, you expect all the traffic between customers to be routed, and you
expect any traffic outgoing from that ISP or *incoming* traffic for
customers of that ISP to be routed to/from wherever they connect to. So if
every node rationally satisifies its customers needs by routing traffic
between its other customers and its connection point(s) to the rest of the
net, then things automatically work out. Again, you pay only for the
connections you make to other nets, and you only charge the customers that
connect to you.
If the industry chooses voluntarily to behave this way, it would be hard to
change, just like it would be hard to buck the internet culture by trying
to use junk mail marketing.
>And their incentive to allow you to
> compete with them by agreeing to allow you to exchange traffic?
>Zero.
Companies have a large incentive to allow traffic to be exhanged. In this
scenario its because it works out automatically, and in general because
that is what is needed to provide value to *their own* customers, to give
them places to connect to, and to build the net. If companies have zero
incentive to agree to exchange traffic, why are they doing it now and why
are they joining the CIX???
