I
want to be careful in responding to a number of these messages. I want to
clarify that the bit price may make indeed sense to carriers. But it is crazy
at as a societal policy and one created by legislation rather than the
marketplace.
Charging
for bits is like Intel charging for CPU cycles instead of selling processors. In
fact in the old days we’d sometimes slow down timesharing systems to
create revenue.
We’re
talking about various accounting and billing models and their consequences including
the incentives. The carriers set the capacity on their networks and they have
an inherent conflict of interest. Having economists set the price is just as foolish
as they are proxies for the carriers since they accept the current business
model.
Before
I get into prices and congestion I’ll note the big “cost” is the
need to have a billing relationship with every provider along the path – the
ability to bill for bits is just a symptom. It means I’m only connected
in locations where I have a prior arrangement. The costs of maintaining the
billing infrastructure are enough in their own right but the real damage is in
preventing ideas that don’t serve ARPU. To put it simply – you’re
disconnected when you leave home. 3G provides some limited respite but is far from
equivalent.
We
should pay for copper, fiber and radios instead of metered bandwidth. As with
CPUs the demand would create capacity. This is the point behind community ownership
of facilities funded as such rather than by the sale of bit transport as a
service.
This should
be simple enough but I’ll go into more detail given the number of
questions raised.
I
agree that infinite bandwidth is an illusion. It works because at any point in
time we have a whole raft of applications that find value in what we have and
that creates an incentive to provide more capacity. This is the virtuous cycle
of hypergrowth. Fortunately we’re already at a point with existing
broadband in that we have a lot of capacity on existing copper and coax as
provisioned.
The problem
is that today capacity is defined by business models and regulation and not by
physics. To keep things simple I’ll use copper as my example but only as
an example – this is not a plea for more copper. Note that copper
capacity has not increased in 25 years because has been no incentive – if
anything the incentive has been to find reasons such as 12000.000000 ft to
claim limits.
Sure
if you move a lot of bits you are taking a portion of the capacity of the
copper wire – the capacity that the carrier has chosen to in order to maximize
ARPU rather than maximizing bit flow. This is one of the perverse aspects of
charging by the bit – it means that increasing capacity would make it
more difficult to justify claims of scarcity. Once you have a gigabit
connection how do you justify a bit charge?
If
you charge for bits then you have every incentive to increase the value of the
bits by limiting them. Not only does a superabundance lead to a price collapse
in such a model, limiting bits means that carriers are get an advantage in
offering their own services. But if the bit business was so good why did Verizon
drop their FiOS plan?
There
is an alternative accounting model in which the community owns the physical
infrastructure and doesn’t pay for redundant “broadband” and
gets 100% instead of 1% of the capacity. If the capacity is used because it has
value then the community would simply add capacity.
From:
nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org
[mailto:nnsquad-bounces+nnsquad=bobf.frankston.com@nnsquad.org] On Behalf Of
Vint Cerf
Sent: Friday, May 07, 2010 16:10
To: Peter Sahlstrom
Cc: nnsquad@nnsquad.org; Matt Larsen - Lists
Subject: [ NNSquad ] Re: Canada goes crazy
resources ARE consumed when you access the Internet. The
access provider has finite capacity to and from the network. To the extent you
are competing with others for a shared access resource, your use prevents
others from having unlimited data rate. So instantaneous bandwidth is the key
resource that is "consumed" second by second.
On Fri, May 7, 2010 at 2:59 PM, Peter Sahlstrom <peter@stormlash.net> wrote:
As has already been pointed out, there is no convenient
analog for
what we are purchasing when we buy internet access, so it's hard to
determine what a truly fair pricing model would be. I think the chief
difficulty is that although no resources are consumed when the network
is used, there is still scarcity in how much network capacity exists
at a given time.
In most cases, ordinary users don't experience bandwidth scarcity any
more than they experience scarcity in electricity or water, despite
the fact that all these systems have limited instantaneous capacity.
An ordinary homeowner can turn on every light and open every faucet in
their house, and they're not likely to see any substantial dimming of
bulbs or loss of water pressure. In order to maintain this illusion
of unlimited capacity, providers must continually forecast expected
loads, and where necessary, upgrade their network to raise the
congestion ceiling. Because these upgrades are driven by the behavior
of the heaviest users during times of peak congestion, it is fair to
expect those users to be the ones to pay the greatest share of the
upgrade cost.
One reasonable model for this type of scaled billing is that used by
the natural gas utility here in Atlanta. Each user's bill lists a
"Base Rate" fee, which includes a "Peaking Service" charge,
which is
calculated annually based on how much gas a customer used during times
of peak demand. This allows the gas company to insure that enough
capacity is available at those peak demand times, and forces the users
most responsible for necessitating upgrades to pay for the costs of
those upgrades. A similar billing model for internet connectivity
would allow service providers to penalize users who cause the most
disruption to the network without imposing limits on how much data can
be transferred, or with whom, or in what way.
I'd appreciate hearing thoughts anyone might have about this approach.
Does this seem fair? Is it practical? Has it already been
implemented anywhere?
--
Peter Sahlstrom
peter@stormlash.net
http://peter.stormlash.net
On Fri, May 7, 2010 at 12:29 PM, Matt Larsen - Lists
<lists@manageisp.com> wrote:
> I'd like to add to the reasons that billing for bits is NOT crazy.
>
> Anyone who has had to manage a network recognizes the need to balance the
> nearly unlimited demands of users with the scarcity of network resources.
> While many of the academics and lawyer types on this list would like
to
> imagine a perfect virtual network that has unlimited capacity, the truth
of
> the matter is that there are limits to a network's capacity and the
network
> resources degrade considerably when users are not limited in how much they
> can consume.
>
> As the operator of a wireless broadband network, I get to deal with
limited
> network capacity on a daily basis. After watching our per-user
bandwidth
> consumption steadily increase, we have begun implementing bit-caps.
A
> detailed description on what we did is available at www.wirelesscowboys.com.
>
> Bits may not be "consumed", but network resources certainly can
be.
> Pretending that network resources are NOT scarce is the the real bad idea.
>
> Matt Larsen
> vistabeam.com
> wirelesscowboys.com
> wispdirectory.com
>
> [ Wireless and wired technologies have fundamental differences that
> can indeed make differing billing and regulatory approaches
> reasonable. What's really problematic is the way
carriers have
> pushed flat-rate data plans in both spheres and now, after
> subscribers are "addicted," want to switch to
measured models.
> One might almost suspect that this was the plan all along,
hmm?
>
> -- Lauren Weinstein
> NNSquad Moderator ]
>
> On 5/7/2010 8:12 AM, Bob Frankston wrote:
>>
>> (Thanks to Aleks for this pointer)
>>
>>
>> http://www.cbc.ca/technology/story/2010/05/06/crtc-usage-based-billing-internet.html?ref=rss
>>
>>
>>
>> The idea of charging people for bits consumed is a crazy idea since
you
>> aren't consuming bits. We've been through this before - do I need to
>> explain
>> once again how bad the idea is?
>>
>>
>>
>> . It creates scarcity. A copper wire (or fiber
or radio) is just
>> sitting there idle. We limit how much can be used.
>>
>> . Even if there is a temporary
constriction somewhere else it
>> means
>> we can't use the capacity locally. To take it to an extreme imagine if
>> there
>> is such a limit in your house - you can't copy too many files between
your
>> computers.
>>
>> . FiOS VoD, for example, goes over IP
through my router. I can't
>> watch much "TV" [sic] if the limit is applied to those bits.
If the limit
>> is
>> not applied we have a vertical playing field where the provider has
all
>> the
>> advantages.
>>
>> . Any sane price doesn't allow making
video affordable if we're
>> going to make the cost of other uses visible.
>>
>> . As with SMS any market that permits
prices to be millions of
>> time
>> cost (determined by competition with Moore's law) isn't really a
market in
>> a
>> useful sense. It's rent taking gone to hostage taking.
>>
>>
>>
>> But basically it shows a deep inability to comprehend the very concept
of
>> connectivity using best efforts. It's railroaders banning the use of
roads
>> unless you buy a ticket for a ride every time you leave your driveway
even
>> if it is just to reorder the cars in the driveway.
>>
>>
>>
>> Others care to add to the reasons why this is crazy?
>>
>>
>>
>> http://frankston.com/public
>>
>>
>> [ And coming soon to a U.S. ISP near you (and me) too,
I'll wager.
>> Since the FCC chairman has shown no interest in
including any
>> sort of pricing or realistically effective
competition-enhancing
>> elements in his proposed "third-way"
regulatory plan, the
>> dominant ISPs are ensured a captive audience of
users who will
>> "pay through their noses until their skulls
are a vacuum" (as one
>> high level ISP executive expressed it to me
yesterday --
>> picturesque, this guy, and a master of invective
as well ...)
>>
>> -- Lauren Weinstein
>> NNSquad Moderator ]
>>
>>
>>
>>
>>
>>
>
>