Re: DHCPv6 clarification draft and PD (Re: [dhcwg] WG last call on draft-ietf-dhc-dhcpv6-opt-prefix-delegation-02.txt)

Jinmei-san,

>>>>>> On Fri, 28 Feb 2003 00:46:37 +0000, 
>>>>>> Ole Troan <ot@cisco.com> said:
>
>>> 4. The PD draft should reflect some parts of
>>> draft-ietf-dhc-dhcpv6-interop-00.txt.  With a quick look, Sections
>>> 1, 2, 3, 4, 5, 6, 9, 10, and 11 should also apply to the PD draft.
>
>> I have made the changes where appropriate, i.e where we already have
>> cut and pasted text from the DHCPv6 base specification.
>
> I don't think it's enough.  For example, Section 8 of
> prefix-delegation-02 is almost the exact copy of Section 22.4 of
> dhcpv6-28, with s/IA_NA/IA_PD/g and s/address/prefix/g.
>
> Since Section 2 of dhcpv6-interop-00 proposes to "add" paragraphs to
> Section 22.4 of dhcpv6-28, corresponding paragraphs should be added to
> Section 8 of prefix-delegation.
>
> I've not gone thorough the entire issues of the interop draft, but I'm
> quite sure that there still exist similar cases.  (If you are not
> convinced, I'll try to make a complete list.)
>
>>> We may be able to omit some of them as trivial clarifications, but
>>> we should reflect some other part of them because the base DHCPv6
>>> spec (and thus the clarifications for it) is too specific to
>>> address assignment.  In some cases, implementors can use analogy of
>>> the base spec to implement the PD draft, but we should basically
>>> provide comprehensive information in the PD draft itself to ensure
>>> better interoperability.  (As some people, including me, have
>>> repeatedly pointed out, the best approach would be to make the base
>>> spec generic so that each stateful method can just refer to the
>>> base spec.  Since we could not make it due to the "it's too late"
>>> reason, we should be responsible to implementors for providing
>>> detailed information within the PD specification).
>
>> the PD specification is not meant to be complete and needs to be read
>> in conjunction with the base DHCP specification.
>
> I know (and agree), but I'm saying the PD specification should be
> clear wherever a difference between address assignment and prefix
> delegation exist.  We should be rather redundant than leave the
> difference ambiguous.  At least please reconsider each issue in the
> interop draft and merge necessary changes from it.

agree. revision 3 of the PD draft should have most of the issues from
the interop document included.

the interoperability testing at Connectathon last week showed that the
DHCPv6 base spec and PD specifications cannot be that bad. no protocol
problems found.

/ot
_______________________________________________
dhcwg mailing list
dhcwg@ietf.org
https://www1.ietf.org/mailman/listinfo/dhcwg