[1093] in Commercialization & Privatization of the Internet
Re: internet consumer reports on state-wide IP networks (fwd)
daemon@ATHENA.MIT.EDU (Mark Seiden)
Fri Jul 26 02:01:13 1991
From: mis@seiden.com (Mark Seiden)
To: bzs@world.std.com (Barry Shein)
Date: Fri, 26 Jul 91 1:31:50 EDT
Cc: com-priv@psi.com
In-Reply-To: <9107251808.AA23834@world.std.com>; from "Barry Shein" at Jul 25, 91 2:08 pm
bzs writes:
>
> mis writes:
> >still confused why a city as big as New York wouldn't justify a full node
> >from very early in the game. Especially when i see the psi POP with
> >its 29 dialup lines, and two expansions so far.
>
> Maybe your answer lies in the second sentance, the icing has been
> licked off the cake (those few easy nodes which would pay for initial
> justification are already on PSI.) But Alternet will be in NYC, I
> think they said the Fall? I really don't know anything more than what
> Andrew said on that particular issue (Doug Kingston might.)
after you posted this i thought you were right, that maybe psi's
sweetheart deal with nysernet has contributed to making them them
hard to beat in new york.
but then i started looking at the usenet maps, and stopped counting after
~40 sites in u.usa.ny.[124] (for some reason i don't have .3) that are
a. currently connected to uunet and
b. within the 5 boros of nyc and
c. not pcs and belonging to companies.
many of these ought to be alternet prospects (given presence and
reasonable pricing) or would be just as happy dialing a local phone
number at a "real node", which would result in some cost recovery.
relatively few mapped new york usenet sites are connected to uupsi
than to uunet, i suspect because uunet's been at it a lot longer.
certainly there's no possible reason given the pricing to talk to
uunet (my experience was >~$100/month for uunet, versus a flat
$25/month for psi, assuming local presence...)
so i don't believe that psi has sucked the jelly out of the doughnut,
the icing off the cake, skimmed off the cream, i love these rich food
metaphors.
> ...
> A lot of your other questions seem to ask for a complete business
> model. It's more complicated than you seem to suspect and the
> questions come off (to me) as sort of naive. For example, there is
> real cooperation between UUNET and Sprint, they're not just another
> telco customer, so overheads are not structured as you represent them,
> and your questions/speculations go off in the wrong direction.
To paraphrase Abernathy, I might be the most qualified person to ask
naive questions about these things...
Are you saying there is no such model, or that they won't say what
their costs are? I would hope they at least have a rule of thumb for
how much demand is needed to justify presence in some area.
Perhaps you would elaborate on the benefits of this "real cooperation"
for the consumer. If uunet has such a nice working relationship with
sprint, would this not reduce the cost for them to set up a "real
node" using Sprint's rack space (versus this virtual circuit thing)?
why is it that uupsi has a host connection on sprintnet/telenet and
uunet doesn't?
Maybe cooperation with sprint reduces uunet's costs, but I only care
if they pass that along. At present, uunet's 800 dialin service is
billed out at .1667/minute at night. Another 800 service I recently
looked at costs .11/minute at night, before any volume discount.
(The only effect of uunet's cooperation with sprint that i ever noticed
was that for months uunet seemed to be diplomatically avoided blaming
either sprint or telebit for a nasty interaction between echo
suppression and trailblazers that would cause excessive retraining and
increase connect time.)
>
> I think the right question is, what does UUNET charge? What do others
> charge (for the same class of service)? Is that reasonable?
that is *almost* the real question. The real question is, what does it
cost the consumer to connect to a carrier, including all additional costs
paid to phone companies or for equipment on customer premises or carrier
premises (e.g. modems and routers).
When one discovers differences in the charge for similar offerings,
it seems only natural to try to track those differences back to cost of
goods sold.
It would be nice if there were spreadsheets to model/compare the
complete costs of the offerings of the carriers at various volumes of
data being carried. It would be instructive for us naive beginners if
someone who's been through it could itemize for comparison purposes
some case studies of the costs of (say) a PSInet and an alternet
connection at 9.6 and at 56k including all those hidden costs.
(parametric descriptions of line costs, e.g. cost per termination plus
cost per mile are also useful...)
>
> Trying to deduce overhead would seem to give one the responsibility to
> do the whole analysis job, not just the parts that are interesting for
> the moment.
>
> (eg. it sounds to my ear, knowing a little more about this situation,
> like the people howling that a SUG/CD only costs $3 to stamp so how
> come...although you do acknowledge the concept of overhead.)
>
I assumed that PSI's and uunet's fixed overheads should be similar, and
was specifically interested in comparing the overheads attributable to
providing a specific competitive service.
(The assumption would be simplistic, if, for example, one company
president has a preference for Porsches and the other doesn't mind
taking the subway, only one has an 800 number, only one has 24-hour
coverage, only one has a salesforce, or if a large amount of one's
overhead is subsidized by a govt contract.)
--
mark seiden, mis@seiden.com, 1-(203) 329 2722 (voice), 1-(203) 322 1566 (fax)
