Re: [DNSOP] Moving forward on draft-ietf-dnsop-private-tld

["Rob Wilton (rwilton)" <[e-mail]>]

Hi Roy, WG,

Roy, just for clarity, am I right to presume that the status of the document that you propose would purely be informational?

It is, of course, up to the WG to decide what to do with this document, but I would like to make a couple of comments that may help the WG.

I would like to somewhat echo a point that was made in DNSOP yesterday when this draft was being discussed, in that I don't believe that IETF should publish a document that either directly or indirectly undermines ISO TC46's ownership or authority over the ISO3166 code points.  I believe that this concern is likely shared by other ADs.

Hence, if the WG decides to progress this document with the proposed structure below, then I'm not convinced that just documenting that these code points exist and that some people use them would be sufficient.  Given the informal liaison feedback that was received, I think that the IETF would likely need to adopt stronger wording that proactively recommends that these country codes are not used for private networks, and highlights the potential problems with doing so.

Regards,
Rob
// Ops AD



-----Original Message-----
From: DNSOP <[e-mail]> On Behalf Of Roy Arends
Sent: 30 July 2021 19:21
To: dnsop <[e-mail]>
Subject: [DNSOP] Moving forward on draft-ietf-dnsop-private-tld

Dear WG

About 40 years ago, give or take, when Jon Postel planned to use the ISO3166 two character code elements as top level domains representing country names, ISO's TC46 secretariat was contacted (as was requested to users of the ISO3166 standard at the time) and he was told that the standard should not be used for DNS, as the future was in X.500. (Postel wasn’t swayed by the argument, and did what we now refer to as permission-less innovation).

Recently, the ISO was contacted again, and subsequently the WG was again told that the standard wasn’t to be used in this way. It seems that a handful of folks are swayed by the argument and want to use this as guidance for the future of draft-ietf-dnsop-private-tld.

Early on, Joe Abley proposed a way forward that I held off initially: Recognise that User Assigned 3166 code elements are used in various ways, including private networks, that these elements have not been delegated and are known to be used to anchor private namespaces. Do not recommend, promote or reserve anything, no registries. Document potential future pitfalls for using these codes for private namespaces and empower readers to make their own decisions.

I now see that with the current status quo, this might a way forward that both sides of the argument might come together on. Essentially, instead of making the pond safe, we’ll have a warning sign that using the pond is at their own risk.

I hope the WG can come together on this as a way forward. 

Warmly,

Roy



_______________________________________________
DNSOP mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/dnsop