[DNSOP] Re: draft-ietf-dnsop-delext and non-useful parent-side types

[Petr Špaček <[e-mail]>]

On 24. 07. 26 13:58, Philip Homburg wrote:
>>> - I assume we will get only one chance to introduce a range of parent
>>>   types. We may want to err on the side of flexibility.
>>
>> I see no reason to assume that.  We can do whatever we think is
>> reasonable.
>>
>> To point out what I would hope is obvious, adding features that
>> nobody use is the sort of thing that makes the DNS Camel cry.
> 
>>From the perspective of a non-recursive DNSSEC validator, it is very hard
> to get such a range deployed. The reason is that upstream recursors and
> forwards have to understand a new range of parent types. We have no
> (secure) signalling for this so the only option is to wait long enough
> that this kind of support is widespread.
> 
> So a future DS2 type will have to use this draft otherwise it will never
> get deployed.

Instead of arguing how complex it could possibly be, I did a very _very_ 
rough "draft of a draft".

I'm sure I made a mistake along the way, but I think it is a good idea 
to get a sense how complex it is.

The magnitude of changes required for having two separate ranges can be 
seen here:
https://github.com/RoyArends/draft-ietf-dnsop-delext/pull/1/changes#diff-900d9177bc56d81d0b08472912da03c2afddd69e36524704baf942c4fc32c513

(It has plain text, Markdown, HTML, and XML formats in it, so pick your 
preferred poison before reading.)

All in all the hardest part seems to be coming up with names for the 
ranges so they can be referenced in the text.

David Blacka proposed these:

*  Parent Types: Combined range of both Delegation Types and Parent
    Information Types defined below.

*  Delegation Types: Designates the set of RR types allocated from
    the ranges reserved in Section 3 of this document.  NS type is not
    a Delegation Type.

*  Parent Information Types: Designates the set of RR types allocated
    from the ranges reserved in Section 3 of this document.  DS RR
    type is similar to types in this range but is an Parent
    Information Type.

The rest of the work is:
- Going through 98 references to "Delegation Types" and changing them to 
one of the above as appropriate.
- IANA Considerations copy&paste.
- New section "Including Parent Information Types in a Referral 
Response", which is almost copy&paste and does the same as we already do 
for DS.

That seems to be it.

I hope this helps the WG to decide.

-- 
Petr Špaček