Re: [v6ops] Making RDNSS a MUST?, hum v2

["STARK, BARBARA H" <[e-mail]>]

> I would have to say that a statement made by an operational group in the
> IETF needs to address the needs of the operators in the group, which I think
> Tore has pretty succinctly stated (and which fits the data arriving in my in-
> box; there are a variety of viewpoints, but if I had to call it right now, I would
> say that the developing consensus is for routers to support both as MTI). If a
> vendor of a product, however widely used, has an opinion, it's relevant and
> welcome (having myself been from the vendor side for quite some time),
> but this is first and foremost about operational viewpoints.

I'd like to share one aspect of my response to v6ops chairs, in response to this. The hum question also asked for thoughts on which document.
What I said about which document was:
 - Leave RFC 7084 alone wrt this. [RFC 7084 already has delivery of DNS by RDNSS and either stateless or stateful DHCPv6 as MTI (MUST) for "CE Routers"; I'm opposed to introducing a DHCPv6 Relay requirement for this class of routers; which doesn't mean they can't or shouldn't implement Relay -- it just means Relay isn't relevant to the purpose of the doc.]
 - Make RDNSS, DHCPv6 Server (stateful), and DHCPv6 Relay MTI (MUST) in the newly adopted draft-ali-ipv6rtr-reqs. I do think that draft needs to be more clearly targeted towards non-mass-market routers (or maybe even just target enterprise, if that's where the big problem is). Many of the requirements are inappropriate for mass market routers (and some should even be recommended against for mass market routers). I'll provide those comments when the next rev comes out.

Note that if we do this, neither document would apply to 3GPP UEs that do tethering.

There is still the separate and completely distinct question about possibly evolving RFC 6434 to make RDNSS a MUST for hosts. Based on my reading of this thread's discussion, I'm leaning more and more towards "Yes" (away from my previous ambivalence, of accepting "Yes" or "do nothing"). 

WRT to the separate and distinct question about evolving RFC 6434 to make DHCPv6 (stateless or stateful) client MTI (MUST) for hosts, I continue to be opposed.
Barbara