[AusNOG] ISP pricing in the NBN world

[Mark Newton]

On Mon, Apr 16, 2012 at 05:21:25PM +1000, Rod Veith wrote:

 > As an ISP, I don't care if you send a resume or a photo of your kid. I do
 > care about existing business models and how I can continue to make money.

But your customers don't care about that.  And they've been enabled
with technology that lets them bypass you if they don't like your
price.

Aside from everything else, the Internet is a marvellously capable
gatekeeper destruction machine.  Doesn't matter what realm you're
talking about, the Internet takes the power previously enjoyed
by gatekeepers, devolves it to end users, and poses the question,
"Now what...?"

That's how it got started in the first place (TCP/IP turned 
circuit-switching telcos into commodity bitpipes, supplanting 
them so completely that they had to turn themselves into packet-
switching ISPs to stay alive).  That's the copyright debate in
a nutshell.  That's the censorship debate.  That's interception
and data retention (the Walsh Report in the 1990's was all about
the question, "What does law enforcement do when the people they
most want to intercept have strong crypto in easy-to-use
packaging?").  That's cloud.  That's Skype and iMessage.

It's all the same theme:  If you behave like a gatekeeper, the 
base technology of the internet provides your customers with the 
means to bypass you.  There is no exclusivity or lock-in associated
with layered services.

Yet layered services seems to be where all the service providers
want to be, now that they've worked out that bit-transport is 
a commodity game with tiny margins.

Look at the new markets service providers want to move into: 
Voice and TV, market domains which significant quantities of
end users have already decided to replace with Skype and Bittorrent.

In the medium to long term, I'd expect those "new markets" to get
commoditized every bit as much as bit delivery has been.  You
won't make money by providing voice telephony to people connected
to your internet service, or pay-TV to people connected to your
internet service.  If you want to make any money at all out of 
those things, you'll need to be offering them to people who
=aren't= connected to your internet service. Become a global player
like Skype and Netflix, where you can afford a yacht and four
Ferraris even if you only get 0.1% of the market.

You can try to lock your internet service customers into your 
layered service offerings with exclusivity.  You could use DPI
to make everyone else's layered service offerings suck (that's
what the Americans do, hence their "network neutrality" debate).
It might even work for a little while.  But deep down you already
know there's no future in it.  The applications and end-users will
find ways around whatever you do to stop them;  the technology
you deploy to make their user experience inferior costs you more
than the profit you can make by selling them TV and voice.

 > As costs come down, competitive pressures will undoubtedly lower the price
 > people pay for services but there always needs to be an adequate margin or
 > the business goes down.

Nobody in Australia gives a crap about whether their ISP's pay-TV
service fails commercially if they can subscribe to another pay-TV
service instead.

You're emotionally invested in your business model because it's
yours.  Your customers aren't. They don't care if you have an 
adequate margin.

 > Of course you can delay paying staff, not pay
 > superannuation, avoid tax and follow other various dubious methods to try
 > and outlast the competition.

You could;  Or you could turn your attention outwards, and ask, "Can 
I sell these services to people in India, the UK and the United 
States?  If not, why not?"

 > There is nothing stopping you offering free voice calls in the NBN world -
 > go for it if you must.

The only thing standing between free calls and paid calls is whether
the SIP handset your customer has bought has a built-in Skype client
and least-cost-routing engine.

It simply isn't your decision.

I'm not talking about this year or next.  It's obvious that there's
*currently* a tidy market for voice telephony, and I can't *currently*
rush out and buy a SIP handset with a built in Skype client and 
least cost routing engine.

(although I can build one myself. Perhaps that's another application
for a Raspberry Pi, eh?)

But the NBN is supposed to be all about the next 50 to 100 years.  
Can you *seriously* see a place for voice carriers in 2062, after
we've brought up several generations of people who have *ONLY* 
known peer-to-peer communication systems, and for whom hub-and-scope
application architectures are quaint inefficient historical foreign
concepts?

How about 2020?

Welcome to the internet.

  - mark