Re: [DNSOP] [art] Another look - draft-ietf-dnsop-attrleaf-05.txt

[Martin Hoffmann <[e-mail]>]

Dave Crocker wrote:
>
> On 3/26/2018 8:18 AM, Martin Hoffmann wrote:
> > Which also reminds me: The DANE RRtypes, ie., TLSA, SMIMEA, and
> > OPENPGPKEY all use underscore labels and are currently missing
> > from the initial table in section 3.1.  
> 
> 
> The table there is for the right-most underscore name, which RFC 6698 
> seems to constrain to choices that are already list.

TLSA needs to be added to the RR column for _tcp, _udp, and _sctp.

RFC 7929 defines _openpgpkey for use with the OPENPGPKEY record type.

RFC 8162 defines _smimecert for use with the SMIME record type.

 
> The left-most underscore name appears to use a rule that differs from 
> the rule used by SRV, etc.  As such, these are second-level names
> being drawn from the same namespace but without coordination.  So it
> looks like there's a problem, but not with the proposed global
> registry (for now.)
> 
> Yes?

Agreed. Currently however, the text in section 2 does sound as if this
sort of collision doesn't happen:

| that is, a subordinate name is meaningful only within the scope of
| the first (top-level) underscore name.

It should probably call out either here or elsewhere that there may
be different hierarchies for different record types.

Kind regards,
Martin