Re: [v6ops] ULA precedence [Thoughts about wider operational input]

[Brian E Carpenter <[e-mail]>]

On 08-May-22 09:53, Michael Richardson wrote:
> 
> Brian E Carpenter <[e-mail]> wrote:
>      mcr> I don't know how it would know to use fd01:abcd:1234:4567::A/64 when
>      mcr> talking
>      mcr> to fd02:fedc:8765::/48 ("ULA2").  Really, this comes back to the general
>      mcr> source address selection problem with multi-homing, which homenet tried
>      mcr> to solve.
> 
>      > Surely, two addresses in fd00::/8 will always have a longer match than
>      > either for them does with any address in 2000::/3 ?
> 
> Sure, but my understanding is that it is proposed to add the /48 associated
> with any fc00::/7 (fd00::/8?) /64 that is found locally.
> That this /48 would now have higher precedence than the IPv4 NAT44 address.
> 
> Think about a laptop/desktop that is inside the company, trying to reach
> smtp.example for outgoing email.  Said name might have a variety of address
> types for inside, outside, VPN, .. maybe even geographically distributed.
> 
>      > That being so, section 10.6 of RFC6724 will still work, won't it?
> 
> Probably.  But, we were discussing prioritizing ULA over NAT44, right?

Yes, but if you prioritize ULA over any IPv4 destination that will always
work, as far as I can see, unless somebody accidentally puts a AAAA record
for a ULA into global DNS.

    Brian