RE: [dhcwg] dhcpv6 'zone suffix' option

> -----Original Message-----
> From: Joe Quanaim [mailto:jdq@lucent.com]
> Send: 2004/11/18 2:11
> To: CTO YAN Renxiang; dhcwg@ietf.org
> Subject: Re: [dhcwg] dhcpv6 'zone suffix' option
> 
> As others have stated, I would prefer using a new option 
> rather than trying to overload the fqdn option.

Yes, as I stated before, using a new option are simple and clean, without adding
complexity to FQDN option.

> CTO YAN Renxiang wrote:
> > For my draft <draft-yan-dhc-dhcpv6-opt-dnszone-01.txt>, the 
> idea behind is
> > rather simple. It's arised when I try to identify a 
> solution to perform DNS
> > auto-registration in IPv6 access network.  and I believe is 
> can also be
> > used in other cases, e.g. DNS update in stateless 
> configuration. Please
> > give me comments, to move this work forward.
> 
> Some comments follow.  I was not at the wg meeting, so I 
> apologize if I am 
> repeating something.
> 
>    Currently, there exist several methods to register domain name for
>    an IPv6 terminal: (1) manually add the DNS resource record into DNS
>    server database; (2) use FQDN option and register domain name
>    automatically either by DHCP client or by DHCP server [6];
>    (3) configure a DNS zone suffix information in the router, and use
>    the RA-based DNS auto-registration mechanism [5].
> 
>    Method (1) is not effective for large number of IPv6 
> devices. Method
>    (3) requires that the every terminal get the address and FQDN using
>    DHCP mechanism. Thus every terminal needs a DHCP client. Method (2)
>    will only be suitable in the case where a router is presented in the
>    network, and the DNS zone suffix has been configured correctly.
> 
> jdq> I believe these references are incorrect.  This  paragraph refers to 
> method 1, 2, and then 3, but it denotes 1, 3, and 2.

It is a typo, I have revised it in the -02 version. Thanks for your mention anyway.

> 
>    DNS zone suffix: A string of DNS zone suffix. It is comprised of a
>    sequence of labels, where each label consists of a length octet
>    followed by that number of octets. The suffix terminates with the
>    zero length octet for the null label of the root. This field should
>    be padded with zeroes to be the multiple of 8 octets.
> 
> jdq> Is it necessary for this field to be a multiple of 8 
> octets?  Also, it is 
> probably better to refer to rfc 3315 than to summarize the 
> dns encoding.  
> This is done in rfc 3646.
> 
>    The above figure shows a typical usage of the option. In 
> this model,
>    ISP has the ISP level domain name suffix (e.g. shtele.com)  The
>    procedure may follow as:
> 
> jdq> It might be simpler to refer to example.com.
> 
>    [5]  Jae-Hoon, J., Byung-Yeob, K., Jung-Soo, P. and Hyong-Jun K.,
>         "IPv6 Router Advertisement based DNS Autocofiguration", draft-
>         jeong-ipv6-ra-dns-autoconf-00.txt, 17 April, 2003.
> 
> jdq> This draft is no longer available on the ietf site.  
> Revision 01 is left 
> as a placeholder saying it has been expired.
> 
>    [6]  M.Stapp, Y. Rekhter, "The DHCPv6 Client FQDN Option", draft-
>         volz-dhc-dhcpv6-fqdn-00.txt, July, 2004.
> 
> jdq> I believe draft-ietf-dhc-dhcpv6-fqdn-00.txt superceded 
> this draft.
> 
> 
> Thanks,
> Joe.

Yes, you are correct, I will make revision in the -02 version. 

Thank you!

-Renxiang