[e2e] Re: economic models for access bandwidth

Henning G. Schulzrinne hgs at cs.columbia.edu
Wed Apr 18 16:33:10 PDT 2001

RJ Atkinson wrote:
> At 15:33 18/04/01, Bob Braden wrote:
> >However, an interesting observation from the discussion so
> >far is that we Internet purists might actually find it useful to have some form of usage-based charging, something we have
> >fiercely resisted in the past.   Useful, because in the real world
> >a provider is likely to try to limit the kinds of service access
> >an end system can have, to limit the bandwidth that user can
> >consume.  If there were charging, there would be no excuse
> >to block arbitrary IP traffic.
> I'd edit the last sentence to read:
>         "If there were charging, bandwidth consumption wouldn't
> be a reason for an ISP to block certain kinds of IP traffic."
>         Also, I think the charging would have to be sensitive
> to offered load, with peak hour bandwidth costing more than
> off hour bandwidth, in order to remove ISP's incentives to
> manage traffic types.
>         Also, residential ISPs cater to the masses,
> who expect their Internet access to be "secure" (whatever that
> means to them).  This sociological phenomenon has led some
> residential ISPs to block certain transport-layer ports
> because they are commonly used to attack the computers
> of the mass subscribers.
>         All that noted, I *really* prefer to have a fixed sized
> and moderate monthly bill for IP dialtone (as I have at present)
> because it makes my monthly budgeting much easier.

I agree, however most people seem to do just fine with variable
electric, bridge toll, gas and phone bills, as long as the variation is
reasonable (you expect to pay more for gas in the winter in NY) and
surprises are avoided. The problem with many telecom charging schemes,
among other things, was that you couldn't conveniently find out until
the end of the month that you had spent far more than planned on, say,
cell phones. (Apparently, it was a common occurrance that people getting
cell phones, prior to 50,000-minutes-a-month deals, would extrapolate
from their home phone bill and then be in for a very rude $500 shock
upon discovering that roaming started at the next city boundary and cost
as much as AirPhone service.)

Gas and electric companies have, for a long time, offered a service
where you pay a "flat" rate each month, based on some kind of
weighted-averaging of your monthly bills, even though the actual monthly
bill varies by a factor of two or more. There's no reason you couldn't
offer something similar for budgeting reasons, even if the actual cost
is usage-based. This would have the right incentives, but avoid most of
the budgeting problems.

> Ran
> rja at inet.org

Henning Schulzrinne   http://www.cs.columbia.edu/~hgs

More information about the end2end-interest mailing list