You appear to be taking the concept to a place where alt is the label
which defines the start of non-DNS switching, and the 2LD is the
specification of the namespace/service you work in. So its a domain
model, driving to software and namespace path, before it begins
resolution of the name proper. (presumably, the remaining F.Q in the
FQDN if DN is now <service>.alt

So where I was asking "why aren't we talking about mechanisms to do
this in omnibar and local control file like nsswitch,conf" you're
saying "given these alternate mechanisms exist, how can we assert a
global name which can identify itself as being addressed to a specific
mechanism for name lookup"

In hypothesis, in this model, ALL FQDN we have "now" are in fact
FQDN.dns.alt ? The dns.alt being assumed, and not necessary in the
default DNS case?

I ask, because NOT grandfathering in the current name scheme into a
new name scheme (as a child/grandchild) invites problems. If the
dns.ALT name is "reserved" then, there can be no confusion in the
mechanisms which do ALT service parsing.

Basically, "if you do .alt, and if the 2lD under alt become a registry
of namespaces/services then you MUST reserve dns.ALT"

-G

_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to