[arin-ppml] Draft Policy ARIN-2019-19: Require IPv6 Before Receiving Section 8 IPv4 Transfers

Michel Py michel at arneill-py.sacramento.ca.us
Wed Nov 6 18:40:57 EST 2019


> I’m quite aware of the report – and I am quoted therein on page 6 arguing a very
> similar point; i.e. that IPv6 may lack sufficient economic incentive to overtake IPv4 - 

I read your RFCs. A long time ago. I was a total zealot at that time, FWIW.

> doesn’t mean that IPv6 is a failure,

Dual-stacking for 40 years is a failure. 20 so far and another 20 to come.
I have been around too, the plan was that we will dual-stack for a small number of years, and that after that IPv6 will become so ubiquitous that IPv4 will die by natural attrition.

I repeat : this plan _has_ failed. And there is no plan B.

What are we looking at today ? 2% or 3% per year ? you call that a success ? 40 years to replace a protocol ?
Reminds me of ISDN : I Still Don't Need. Look around you and tell me where you see ISDN. There is still plenty of POTS, and plenty of aDSL over POTS, not over ISDN.

> However, if you wish to interconnect with others,

None of my customers have IPv6. None of my suppliers have IPv6. My current upstream does not have IPv6. If Google goes IPv6-only, I will find another search engine that values my business. There are ZERO others that I wish to connect to that require me to install IPv6. I have ZERO need for IPv6 and I won't have any for the foreseeable future. The only thing that IPv6 could do for me is to waste my resources configuring it.

It's been 20 years, and you need to tell me that I need to dual-stack and wait another 20 years ?
You want me to go tell my investors that I need money to spend on an upgrade that provides ZERO ROI for at least 10 years ?
That brings me no new customers ? that does not reduce any of my OPEX ?

> since the coordination that you benefit from in terms of registry-provided
> uniqueness has a corresponding cost of compliance with the registry policy.

If such policy comes to pass, it will be challenged. ARIN has no business forcing me to spend time to configure IPv6 in order for me to administer IPv4 resources.

This proposal is flawed. Forcing people to register IPv6 prefixes so the IPv6 deployment figures look better. It's a lie.
It's a lie to make some numbers look good.


More information about the ARIN-PPML mailing list