Re: Last Call: <draft-iab-2870bis-01.txt> (DNS Root Name Service Protocol and Deployment Requirements) to Best Current Practice

[David Conrad <[e-mail]>]

Russ,

On May 30, 2014, at 11:30 AM, Russ Housley <[e-mail]> wrote:
> I think it is important for the IETF to say that each of the root servers -- all 13 of them -- need to support IPv4 and IPv6.  

Assuming root servers are "IP-Capable Nodes", the IETF has already done that via BCP177 and yet 3 root servers still do not provide IPv6 service. I fail to see how yet another document is going to change this, particularly when you have two root server operators either saying explicitly or implying that the proposed BCP will have no impact.

> The root servers are critical infrastructure, and in my view, they should lead the way to IPv6.

The main impedance mismatch I'm having here is that in the vast majority of other cases, a document published by the IETF saying "Thou Shalt" can be used by folks to go to vendors or service providers and say "if you do not obey the IETF's Thou Shalts, I shall take my money elsewhere". This obviously does not apply to the root server system or individual root server operators. As such, I feel 2119 language is just silly, particularly given the root server operators presumably already know the community wants them to do stuff like support IPv6.

If this is just a "feel good" exercise like some odd Twitter hashtag campaign, then I suppose publishing the document probably won't hurt.  I just think it'd be a mistake to assume it actually does anything.

Regards,
-drc