Re: RIPE our of IPv4

[Brian Knight <ml () knight-networks com>]

> > On Nov 27, 2019, at 2:54 PM, Brandon Butterworth <brandon () rd bbc co uk> wrote:
> >
> > On Wed Nov 27, 2019 at 01:08:04PM -0600, Brian Knight wrote:
> > None of which matters a damn to almost all of my business eyeball
> > customers.  They can still get from our network to 100% of all Internet
> > content & services via IPv4 in 2019.  I regularly vet deals for our
> > sales team, and out of the hundreds of deals we sold this year, I can
> > count on one hand the number of deals where customers wanted IPv6.  We
> > sold them IPv6 access, but we didn't put it on our own network, because
> > we face the same internal challenges Sabri mentioned.  (SD-WAN, OTOH,
> > was far more popular
>
> A few year later customer wakes up:
>
> "wait you sold us all those toys we didn't need but didn't include
> the basic transport capabilites everyone apparently has been saying
> for over a decade are required minimum?"
>
> "and now you want us to pay you to rebuild it again and trust that
> you got the basics right this time?"
>
> If you're an internet professional you are a negligent one if by
> now you are not ensuring all you build quietly includes IPv6, no
> customer should need to know to ask for it. It's not like it
> needs different kit.

Possibly some customers may react this way, but I’m thinking many more would ask “what does it take to enable it?”  
Most are reasonable and show good faith, even if an equipment swap is needed.  And if the demand for IPv6 is there, the 
providers will get the work prioritized.

> > As an aside, I would guess that it's the corporate eyeball customers
> > with servers, not resi/mobile behind CGNAT, that will bear the brunt of
> > the IPv4 cost first.  But what enterprise wants to tell its non-IPv6
> > customers "your Internet needs to be upgraded, come back to us when
> > you're done?"  That doesn't bode well for the short-term future.
>
> "all that multi natted into same address space VPN firewall 
> complicated knitting we never got right wasn't needed if you'd
> told us to use IPv6?"

IPv6 doesn’t help anyone get access to their IPv4-only customers.  (Too bad that it doesn’t.)

My point was that, if eyeball networks start charging a premium for IPv4, their likely first customers to be charged 
are business customers not behind CGNAT.  Those that don’t wish to pay the IPv4 premium would have to force *their* 
customers to go IPv6. That would be a much more difficult conversation than simply paying the premium.  So out of all 
the forces at work, which gives way first?

> brandon

Thanks,

-Brian