[lacnog] Fwd: Re: Usage of services without IPv6 Support

[Fernando Gont]

Buen dia,

Thread interesante en la lista general de IETF (ietf en ietf.org)....


-------- Forwarded Message --------
Subject: Re: Usage of services without IPv6 Support
Date: Wed, 22 Apr 2020 15:03:24 +0200
From: otroan en employees.org
To: Keith Moore <moore en network-heretics.com>
CC: ietf en ietf.org

>> We are 20+ years in. If we (as the collective we) believed in this
>> transition shouldn't the next step be to remove the A records from
>> {mail,datatracker,www}.ietf.org? And add a piece of Javascript
>> blocking any client coming through a translator from v4.
> Our recommended transition strategy was dual-stack.    Removing IPv4
> support is premature until IPv6 is ubiquitous.   Nobody is proposing
> that github be made IPv6-only.
>> I would expect ridicule from such a proposal. Our tools including
>> our collboration tools are centralized, they don't benefit much
>> from IPv6 and end to end transparency either.
> 
> No but it's entirely possible that there are, or soon will be, IETF
> participants who cannot access github because of the lack of IPv6
> support.   As IPv6 becomes more available and IPv4 space more and
> more scarce we should expect that some existing access networks will
> discontinue IPv4 support.   Who wants to maintain equipment and
> routes that aren't used anymore?   When public IPv4 network access
> disappears, it will happen gradually at first, and then everywhere
> suddenly.
> 
> I don't think the github situation is an emergency yet, but I agree
> that we need to make sure that all of the tools that we use and
> expect participants to use are accessible via IPv6.  Maybe not this
> week, but probably this year.

I'm quite aware of the IETF transition strategy.

20+ years in isn't apparent that it is working. The IETF cannot disable
IPv4 from it's properties in the forseeable future. It is equally
unforseeable when a client can communicate only IPv6 end to end.

The sorry reality is that the centralized network combined with the fact
that a single IPv4 address is sufficient for a small country (with a
port overloading NAT) there is no urgency to the transition.

I would love to be proven wrong of course...

Cheers,
Ole