[v6ops] Re: [E] New Version Notification for draft-mishra-v6ops-variable-iids-problem-statement-01.txt

[Lorenzo Colitti <[e-mail]>]

On Thu, Sep 19, 2024 at 8:27 PM Ivan Pepelnjak <[e-mail]> wrote:

> If there is incentive for ISPs to move from /56 to /80 (e.g. due to
>>
> ARIN fee structure) this incentive should be corrected where it comes
>> from.
>>
>
> Are you asserting that ISPs that provide only a /64 today are *not* going
> to move to /80 once /80 becomes widely accepted?
>
>
> So what? You are able to deploy one SLAAC-only subnet with direct Internet
> connectivity when an ISP gives you a /64 today… unless you have hosts that
> run DHCPv6 clients (most platforms these days with a few rare exceptions),
> in which case you can have as many subnets as you wish today, regardless of
> what the ISP is giving you.
>
> What’s the difference if the minimum widely accepted SLAAC-only subnet
> size is /80 tomorrow, and the same ISP gives you /80?
>

As you point out, that is the likely outcome: a race to the bottom.


> One example is Comcast. Another example: in most of Japan, fiber ISPs only
> provide a /64, and to get a /56 you have to sign up for a separate service
> that costs the equivalent of $5/month.
>
>
> So you’re saying that your personal disgust with a particular ISP should
> shape the development of IPv6?
>

I'm not blaming the ISPs for responding to incentives. I'm just saying that
they will continue to do so, and that we must take that into account. If we
change the minimum acceptable subnet size to /80, then we incentivize those
ISPs to use /80 instead of /64. Some will do so. As soon as that happens,
/80 will become the lowest-common denominator. Soon after that, the problem
Gyan is currently talking about ("oh dear I only have a /64, I need to
subnet further but I can't") will present itself again, except with
s/64/80/. In other words, we will have a race to the bottom. That is a bad
outcome and we should avoid it.