GTE BBN - Rand originated BUSINESS MODEL discussion
rirving at onecall.net
Sat Aug 22 17:22:42 UTC 1998
You cannot configure your router from this message... warning.
This is continuation of the "BLUE SKY INTERNET BUSINESS MODEL"
Obviously I am missing something..( I may be naive on this)
Why would you want to charge hosters for hits ? (seriously !)
I cannot conceive charging an author of a popular book
for everyone who reads a copy.... Most could never
afford it.... Worse, popularity would create hardhship.
Now streaming UDP, (UNI/MULTI) growing in popularity so
rapidly for multimedia distribution, could be originator
based. More of a Television Industry method... The origin
would want to make sure they have the credit card of the
user who is about to tune into their latest
"multimedia transmission" before they start the
*push* based application. (Back to "no surprises" concept)
Also, to someone else's point...
(in a *much* more *real world now* discussion :)
In my *opinion*:
Your right, details of peering should be between peers. It already
Next "Lata" (Peer AS's ?) transactions should go to the owner of the
(Which if they pull to you, is them...)
(If you pull to them, it is you... If you choose not to charge, your
(Bonus: Owner of last physical collects the *hypothetical* acks fee's.)
However, in my heart, I do not feel that the users are *not*
They performed an "ACTION" to cause that data to flow... The hoster
did *not*...... Remember, NO surprise bills!
I really feel that this is a better philosophy.. I would hate someone
being able to call
my grandmother collect, without her having the option to say *NO*....
Also, popularity *should be* *desirable* on the internet. But, lets
at the economics of it... THE WORLD could decide to descend on
"IGGY the IGUANA" ( a simple mid-west families site hosted on our
*extremely popular* with children worldwide..)
Now, should that family be billed 1 million at the end of the month ?
Or, should the provider be billed 1 million.
Which of these should be punished for providing interesting content ?
Or , how about, the 1 billion users who went there ?? What if they were
billed about .001 cents for that transaction, would they be discouraged?
I Don't think so...
You billed the same 1 million dollars, but everybody paid a price about
equal to the service's they received. (and with the hypothetical ack
fee's, the ISP could afford to put up pipe to support IGGY's popularity)
("Morris the Cat" sure never "hurt" for cat food :\
Regarding complexity of DS0m/s billing.....
Certainly this could be approached with a "region uniform pricing"
to the end users...
"Surf AnyWhere in the Continental North America, for
a simple flat fee .0004c DS0/s! With "Sprat Cents" no more confusing
(Any relationship to a real companies approach is purely hysterical,
er.. hypothetical ;)
But, that wouldn't mean that the *providers* wouldn't settle up in a
controlled detailed fashion... (or maybe not, interesting....)
"Cross MCI pathways for a simple .0001 cents a DS0/second, good anywhere
in continental america, and your MCI local region remains free"
Finally, a fiscal motive to peer.
Also, given a hypothetical unit DS0mile/seconds.. What would prevent
someone from routing poorly to increase the "cost of call", as the
throughput" (which in turn would generate volume in terms of DS0 *mile*
seconds) does not
go down *that* linearly with distance... maybe you could throw in a
Mile bands are based on *geography*, not physical path taken. ;)
Or even upon "regions" (AS's with SUB-'AS's,
a model already currently in use [check your AS-SETS], and regional
This sure could motivate inter-provider meshing..... And assist in
"across the street, via MAE-E" routing...
The more efficiently your route your data, the greater the margin, as
are reduced... Less efficient ? Your loss, higher cost of call.. You
(In terms of reduced margins)
Now data loss becomes an issue.. This unfortunately would be a
as it is the pipe of the ISP connected to a "poor quality" NSP that will
be consumed by lossy
fabric. Perhaps the accumulated CDR would result in fiscal punishment to
the lossy NSP..
"5000 DS0/s were delivered to said NSP, 4000 DS0/s were completed
150000 DS0m/s".. collected.
Hrmmmmm.. A reason to change NSP's ? After all, that NSP is losing you
NSP's who have a greater completion rate would be a competitive angle...
Better circuit, better management = greater DSO/s completion ratio =
"Use NSP SOMEPROVIDER. We guarantee a 98% DS0/s completion ratio, if
DS0/s completion ratio drops below this threshold we will pay you
small print follows, DS0 Second payments based on average "DS0 miles"
on existing completed seconds of traffic, UDP exclusion clause, etc,
Maybe I should get some coffee.... I am taking this too seriously.
More information about the NANOG