> On Nov 19, 2021, at 07:23 , Dave Taht <dave.t...@gmail.com> wrote:
>
> On Fri, Nov 19, 2021 at 7:00 AM Owen DeLong via NANOG <nanog@nanog.org> wrote:
>> Since, as you point out, use of the other addresses in 127.0.0.0/8 is not
>> particularly widespread, having a prefix
>> dedicated to that purpose globally vs. allowing each site that cares to
>> choose their own doesn’t seem like the best
>> tradeoff.
>
> I would prefer to discuss the other drafts. However, - and this is not
> in the 127 draft, and is an opinion not shared with the other authors
> -
> I have a specific use case for making 127 "more routable", in that
> there is nowadays a twisty maze of microservices, bottled up in a
> variety of
> kubernetes containers, running on top of vms, on top of a hypervisor,
> that are often hooked together via rfc1918 addressing and NAT.
>
> Trying to figure out that particular path, from within one of those
> containers, can be a PITA. The concept of 127 being local to a
> physical host
> (and routed internally, rather than natted), where those twisty maze
> of services ideally remains within that host, holds some appeal to me.
Couldn’t you do this with 169.254.0.0/16 (or IPv6 GUA or ULA)
just as easily without the need to rewrite virtually every layer of the
stack of miscellaneous software defined networking stuff you just listed?
Owen