North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

RE: GTE BBN - Rand originated BUSINESS MODEL discussion (fwd)

  • From: Terence C. Giufre-Sweetser
  • Date: Wed Aug 26 11:33:20 1998

On Sat, 22 Aug 1998, Mike Leber wrote:

> It sounds like you are proposing settlements based on subtracting received
> from transmitted (or dialup provider pays, web hosting company earns). 
// snipped heavily...

Hang on, the main revenue stream for a dialup isp is always user's 
subscriptions and fees.  A good web site makes money from plain old 

What are we selling here?  A network access product? Or bytes of useful 

Network access implies we're selling the use of the pipes, and big 
content providers pay big money to get their content into that network, 
the same way you pay for a pipe running off the best connected network, 
and users pay dailup subscriptions to have their 33 to 56 kbps.

Bytes?  We're buying the stuff IN the pipes, the network no longer 
matters, it's just a way of getting bytes here and there, run it only for 
the sake of getting bytes pushed.  Now, do you pay for bytes to be sent, 
or do you pay for bytes to be received?  Well, it's mostly consumer 
based, so the receiver pays.

Imagine if the US postal service are just like a byte payer internet.
You walk down to the post office and mail your letter, it gets handled by 
many people (routers) and sorting systems (more routers) and crosses 
through quite a few different companies, (as it's the internet, no one 
entity owns all of the net,) and these companies all take a fee for their 
service, so, you pay 40c for the stamp and that 40c is whittled down to 
nothing by the time the final destination is reached.

Now, that may work well, but the model ignores Quality of Service, and 
the capacity to carry lots of mail in special heavy use periods.  The 
profitability of the service is suspect, because a competitor may come 
out with a 35c service and lay your service (network) idle and 
unprofitable.  Service between particular points may be suspect, however 
these flows of mail are moved to the bigger mail handlers, making the 
marketplace heavily stratified, with anti-competitive practices of the 
big mail handlers forcing the little guys to use them as go-betweens.

So, in total ignorance of mail flows (network topology) the market suffers
bloated pricing due to stratification of the operators, no guarantee of mail
arriving on time (there's no bandwidth guarantee) and of course, political
fighting between mail handlers over who should be a customer paying for
delivery and who should be taking the biggest cut of the 40c or 35c. 

Look at the carrier/courier market and you have this right now.  People 
are making money, some are going broke, and common carriers are making 
life difficult for smaller operations and regional operators.

The analogy falls apart when you realise that the US postal service and
DHL et al do not really care what they cart, they'll deliver empty boxes
if you pay them, and so, the whole point is that no one really knows how
much real "bandwidth" is there out, whether or not you get to use it,
and of course, no one really cares what you're shipping...