Re: [IPv6] Second Working Group Last Call for <draft-ietf-6man-rfc6724-update>

[Jared Mauch <[e-mail]>]

> On Apr 15, 2024, at 2:49 PM, David Farmer <farmer=[e-mail]> wrote:
> 
> You say, "In these cases, we will get the old behavior, which works well enough."
> 
> However, RFC6724 is not implemented universally or entirely. Therefore, the current behavior, or the old behavior compared to this update, is a mix of RFC3484, RFC6724, and a hybrid of the two in gai.conf. 
> 
> Moreover, you and others might feel RFC6724's behavior is good enough, but at least many of those instigating this draft think preferring IPv4-IPv4 or ULA-ULA is incorrect behavior and contrary to the intent of preferring IPv6 to IPv4.
> 
> It would be best for all IPv6 implementations to implement known-local ULAs. The normative language best suited to accomplishing that is debatable; I'm fine with SHOULD or MUST, with a slight preference for MUST.
> 
> Nevertheless, any implementation that does not implement known-local ULAs should raise the ULA label's preference above that of the IPv4 label. Give that at least most implementations that use gai.conf already do that that is not inconsistent with the current behavior works good enough.

It seems like if the goal is to tell the host to modify the ULA behavior the right way is likely to allocate a bit via PIO, but hosts that receive a static config, it would be incumbent upon them to modify their local behavior and can’t take a remote hint.

- Jared