Re: [DNSOP] Working Group Last Call for draft-ietf-dnsop-rfc6598-rfc6303-00

[Suzanne Woolf <[e-mail]>]

David,

I'm glad you spotted this. I think it is indeed a concern.

On Dec 7, 2013, at 9:09 PM, David Conrad <[e-mail]> wrote:

> I've read the document and I think there might be a bit of an issue:
> 
> It appears we now have multiple registries listing the same information.  Specifically:
> 
> http://www.iana.org/assignments/locally-served-dns-zones/locally-served-dns-zones.xhtml
> 
> seems to be a subset of
> 
> http://www.iana.org/assignments/special-use-domain-names/special-use-domain-names.xhtml
> 
> Options to deal with this:
> 1. revise the IANA considerations section to indicate that both registries need to be updated
> 2. remove the locally served zones from the special-use-domain-names registry
> 3. merge the two registries into a single uber-wonky-name registry
> 4. give up on this DNS thing as it was clearly a mistake and go back to ip address literals
> 
> While the last option is looking more and more appealing, I'm personally leaning towards option 3, adding a column that says how the name should be dealt with (e.g., "locally served via DNS", "never touch the wire", "non-DNS", etc.).  

Which answer makes the most sense depends a bit on whether there's any urgency to getting this draft published.

I suspect that merging the two registries, including agreement on a set of possible values for "what to do about it", is the right answer but also likely to take longer than having this draft update both existing registries (either of the remaining options you propose). 

I'd support your option 3 if there's no urgency here (and I'm aware of none). 

If people are in a hurry for some reason, option 1 seems marginally less likely to lead to confusion than option 2 (neither registry that people might be relying on today gets *less* useful), but it seems clear to me that somewhat idiosyncratic partial overlap between two registries is something we should avoid where possible.


Suzanne