Some truth about Comcast - WikiLeaks style

Thu Dec 16 15:17:04 UTC 2010

I disagree with this theory.

If customers pay comcast for bytes then eventually the upstream (L3) will
want some of that revenue. That revenue will be passed onto the provider as
a lower bill.

This encourages Netflix to send more bytes, because if they do Comcast and
L3 get paid more and Netflix's bill goes down. The income stream is now
completely dependent on how much money the customer can pay. Sure other
services (Hulu, Youtube, etc) could step in and offer lower usage but
customers are MUCH more likely to pick based on available content then an
abstract "we use N bytes per month fewer then that guy". Also, if income is
dependent on sending more bytes, provider usage will creep up as far as they
can get push it while still keeping customers.

In comparison, does the average person pick their mobile provider based on
how many minutes they get or if they get coverage at their house. Keep in
mind that all providers are within ~10% of each other on pricing. Where do
we see costs in mobile going? How much is a text message these days? That is
what happens when the user pays for bytes.

If the provider pays for transport then the amount of data going over wires
is dependent on that providers customer base. The cost of transport now
scales directly with size of business. Netflix or Hulu are now directly
responsible for their costs which motivates them to be efficient.

Comcast can now charge its customers only for upkeep of its network and use
the income they get as an "end point delivery network" to offset customer
cost. Comcast's cost, which are upkeep and expansion of its physical
network, now scale proportionally with its customer base.

So in this model, customers pay for the laying of the wire to their house
and the upkeep of that wire, which is a 1:1 ratio for the consumer/comcast.
Providers (Netflix) pay per byte to the transport providers. Transport
providers make thier money off providing transport.

Income and cost of doing business for Netflix is directly tied to their
subscriber base so it is easy for them to balance their income.

Of course, I am not an economist and could be entirely wrong. There are
certainly other HUGE political factors, but I think in theory we would all
be happier if the system worked by someone paying for a postage stamp then
COD.

On Thu, Dec 16, 2010 at 1:05 AM, George Bonser <gbonser at seven.com> wrote:

>
>
> > From: JC Dill
> > Sent: Wednesday, December 15, 2010 9:13 PM
> > Cc: nanog at nanog.org
> > Subject: Re: Some truth about Comcast - WikiLeaks style
> >
> > Sure, Comcast's customers are also paying Comcast.  But Comcast wants
> > to
> > get paid from the content provider.  I think they are betting that in
> > the long run it's easier to make money from content providers (and
> have
> > the content providers charge customers or advertisers as necessary to
> > make a profit) than to make money from the end consumer.  And I think
> > they are right about this "easier" part.  I think that they will
> > succeed
> > at pressuring big content providers to play by Comcast's rules and
> > shift
> > the cost of running Comcast's network from consumers to content
> > providers.
> >
> > jc
> >
>
> There are two different innovation paths according to who is paying.  If
> the customer is paying, innovation is driven by the interest of the
> customer.  If the provider is paying, innovation is driven by the
> interest of the provider.
>
> If the customer pays the cost of the transport, a provider with better
> transport efficiency / quality ratio wins.  It spurs innovation where we
> get better quality product with a better transport efficiency.  If there
> are three competing content services in the market offering basically
> the same quality product, the one with the better transport efficiency
> is going to win customers.  Or in some cases the customer might choose
> to sacrifice some quality for transport efficiency.  The market
> eventually settles on what the customers in the aggregate decide is
> their willingness to trade price for performance.
>
> If the provider pays the cost of the transport, a provider might
> effectively subsidize the transport cost of a bloated content
> distribution mechanism.  It won't make any difference to the last mile
> delivery network either way.  Either way they get the same amount of
> money.  If provider pays the freight, there might be some company with
> an absolutely killer technology that can stream much higher quality
> stuff with less bandwidth usage but if the customer doesn't see the
> benefit, that in and of itself isn't enough to drive eyeballs to that
> content.  If that content transport method did save the customer money,
> the eyeballs would move in that direction.
>
> Having the provider pay the cost stifles technological advancement.  It
> facilitates a "deep pocket" established company creating a barrier of
> adoption to a startup who might have a more efficient product but the
> user doesn't get any direct benefit so they don't adopt it.  Having the
> user pay gives an incentive to develop technologies that reduce the
> network burden.  Having the provider pay distorts innovation.
>
> In the end, having the end user pay the cost for the product they are
> consuming results in better, faster, cheaper (yes, you can have all
> three).  Externalizing those costs through subsidies by outside parties
> throws things out of balance and drives innovation in a way that
> benefits the provider, not the consumer.
>
>
>
>