RE: CPE router learning DNS servers - comment on draft-ietf-v6ops-ipv6-cpe-router-03

["Dan Wing" <dwing@cisco.com>]

Thanks!
-d
 

> -----Original Message-----
> From: Ole Troan [mailto:ot@cisco.com] 
> Sent: Thursday, January 14, 2010 3:36 AM
> To: Dan Wing
> Cc: 'STARK, BARBARA H (ATTLABS)'; 
> draft-ietf-v6ops-ipv6-cpe-router@tools.ietf.org; v6ops@ops.ietf.org
> Subject: Re: CPE router learning DNS servers - comment on 
> draft-ietf-v6ops-ipv6-cpe-router-03
> 
> Dan,
> 
> to make sure this has been answered.
> 
> >> The L-6 requirement is a LAN-side (DHCPv6 Server) 
> >> requirement.
> > 
> > You're right; my misread.
> > 
> >> However, since SLAAC is the expected method for address config in 
> >> the LAN, and support for DHCPv6 DNS_SERVERS option is a MUST (and 
> >> other options is a SHOULD), I think that the MAY statement for 
> >> stateless DHCPv6 support might not be completely consistent.
> > 
> > Agreed - we certainly need a way for the hosts to learn their
> > IPv6 DNS servers.  And it should agree with RFC4339 and RFC5505
> > (neither of which are cited by draft-ietf-v6ops-ipv6-cpe-router-03).
> 
> these changes should hopefully clear the inconsistency:
> 
>    L-10:  The IPv6 CE router MUST support a DHCPv6 server capable of
>           IPv6 address assignment according to [RFC3315] OR a 
> stateless
>           DHCPv6 server according to [RFC3736] on its LAN interfaces.
> 
>    L-12:  The IPv6 CE router MUST support providing DNS information in
>           the DHCPv6 DNS_SERVERS option [RFC3646].
> 
> 
> with regards to RFC4339, I consider that historic at this point.
> as for RFC5505 I see that more as guidelines for protocol 
> developers. we are trying hard to invent nothing in this 
> draft, so I'm not sure it is relevant to cite 5505 either.
> 
> >> As for the WAN-side (DHCPv6 client), there's currently no 
> mention of
> >> stateless DHCPv6. It does seem implied by other requirements that
> >> describe support for SLAAC + the ability to get other 
> config info via
> >> DHCPv6. It might be good to be more explicit on this.
> 
> since we have a MUST on a stateful DHCP exchange on the WAN 
> interface I don't see the purpose of mentioning  the 
> so-called stateless variant.
> 
> cheers,
> Ole
>