[IPv6]Re: [v6ops] New draft: "The IPv6 Loopback Address Prefix"

[David Farmer <[e-mail]>]

Geoff,

Ok, I see where you discuss IPv4-Mapped IPv6 Address, ::ffff:0:0/96, but I
don't see a discussion of the deprecated IPv4-Compatible IPv6 Address,
::0/96. Yes, plenty of time has gone by, but we should at least acknowledge
the reuse. Furthermore, do we really need 4B loopback addresses? Isn't 16M
enough? Or 64K or 256 for that matter?

Thanks.

On Tue, Nov 25, 2025 at 4:41 PM Geoff Huston <[e-mail]> wrote:

> Hi David,
>
> I had noted that, but also noted that RFC4291 is now twenty years old! If
> twenty years is not a long enough period to treat a deprecation action as
> "concluded" then we are in far FAR more trouble with the IETF standards
> process than just the lack of an IPv6 loopback prefix!
>
> Given that the text of RFC4291 clearly states in Section 2.5.5.1
> (IPv4-Compatible IPv6 Address) that "The "IPv4-Compatible IPv6 address"
> is now deprecated" I did not see any point to further twiddle with this
> part of RFC4291 in this draft.
>
> So, no, I obviously do not concur with your view this is the "kiss of
> death" for the concept of a loopack IPv6 asddress prefix at the start of
> the IPv6 address set.
>
> You also commented:
>
> Note that if we go hard with this idea, the entire IPv6 loopback prefix
>> should be added to the Locally-Served DNS Zones registry.
>>
>>
>> https://www.iana.org/assignments/locally-served-dns-zones/locally-served-dns-zones.xhtml#ipv6
>>
>
>
> That' a great suggestion David - will be added to the draft
>
> Thanks,
>
>   Geoff
>
>
>
> On 26 Nov 2025, at 9:16 am, David Farmer <[e-mail]> wrote:
>
> Also, I think there should be a discussion of how this overlaps with the
> now-deprecated IPv4-Compatible IPv6 Addresses in section 2.5.5.1 of
> RFC4291. In fact, it might be better to use ::/104, which maps to the IPv4
> prefix 0.0.0.0/8.
>
> This is probably the kiss of death for this idea, but to avoid confusion,
> it might be better to produce a full RFC 4291bis and eliminate section
> 2.5.5.1 of RFC 4291. However, the last time we tried to do an RFC 4291bis,
> it ended in total failure.
>
> Thanks
>
>
> On Tue, Nov 25, 2025 at 2:31 PM David Farmer <[e-mail]> wrote:
>
>> While I'm not strongly opposed to this idea, I have always liked that the
>> IPv6 loopback address is a singular address rather than a prefix, as in
>> IPv4.
>>
>> I don't think it is a major issue, and there are more than enough IPv6
>> addresses to define a lookback prefix for IPv6, instead of a singular
>> loopback address. It won't cause a shortage by any means.
>>
>> I guess it is mostly a matter of taste. However, I would prefer a better
>> justification than simply saying we did it that way in IPv4, which seems to
>> be the primary justification.
>>
>> Note that if we go hard with this idea, the entire IPv6 loopback prefix
>> should be added to the Locally-Served DNS Zones registry.
>>
>>
>> https://www.iana.org/assignments/locally-served-dns-zones/locally-served-dns-zones.xhtml#ipv6
>>
>>
>

-- 
===============================================
David Farmer               Email:[e-mail]
Networking & Telecommunication Services
Office of Information Technology
University of Minnesota
2218 University Ave SE        Phone: [phone]
Minneapolis, MN 55414-3029   Cell: 612-812-9952
===============================================