Re: [Mip6] Questions ondraft-ietf-mip6-bootstrapping-integrated-dhc-00.txt

Junghoon Jee wrote:
> Hi Alex, Let me try to reply to your questions.
> 
> ----- Original Message ----- From: "Alexandru Petrescu" 
> <alexandru.petrescu@motorola.com> To: <mip6@ietf.org> Sent: Saturday,
>  April 29, 2006 1:29 AM Subject: [Mip6] Questions 
> ondraft-ietf-mip6-bootstrapping-integrated-dhc-00.txt
> 
> 
>> Dear authors of 
>> draft-ietf-mip6-bootstrapping-integrated-dhc-00.txt, I have some 
>> questions and comments on this draft.
>> 
>> Shouldn't the DHCP options defined in this draft be proposed to the
>>  DHC WG and not to the MIP6 WG?
> 
> There's already a draft for that purpose: 
> http://www.ietf.org/internet-drafts/draft-jang-dhc-haopt-02.txt
> 
>>> [fallback case] In the fallback case, the mobile node is not able
>>>  to acquire the home agent information via DHCPv6.  The mobile 
>>> node performs DNS queries to discover the home agent address as 
>>> defined in [BOOT-SPLIT].  To perform DNS based home agent 
>>> discovery, the mobile node needs to know the DNS server address. 
>>> How the mobile node knows the DNS server address is outside the 
>>> scope of this document.
>> What are the conditions in which the fallback case occurs?  Is it 
>> because DHCP does not respond?  Or just that the DHCP responds with
>>  an error when queried about HA?
> 
> If I understood your question correctly, the fallback case here means
>  where the integrated scenario does not apply,

Nowhere in the draft-integrated does it say that the fallback case
happens that the integrated scenario does not apply.

> more specifically ASA != MSA.

How does the MN know that ASA!=MSA?

>> If the latter then the DNS address could be still obtained from 
>> DHCP. Otherwise, the DNS address can be pre-configured (in 
>> /etc/resolv.conf) or from a LCP (link-control protocol).  I don't 
>> understand why it's left outside the scope of this document.
>> 
>>> (2) The mobile node sends a DHCPv6 Information Request message 
>>> [RFC3315] to the All_DHCP_Relay_Agents_and_Servers multicast 
>>> address. In this message the mobile node (DHCP client) SHALL 
>>> include the Option Code for Home Network Identifier Option 
>>> [HAOPT] in the OPTION_ORO, Home Network Identifier Option with 
>>> id-type set to 1 and the Home Network Identifier field set to the
>>>  network realm of the home MSP [HAOPT].
>> Currently a network realm part of NAI designates several subnets 
>> (HAOPT uses NAI).  So if in a network domain (realm) there are 
>> several subnets, there should be several home agents there.  How 
>> does DHCP and AAAH know which HA to return to MN when the 
>> identifier is only a generic network realm?
> 
> Not so sure if I understood your question here: Anyway, MSP/AAAH 
> assigns a HA in the integrated scenario.

Yes, it does.  But for draft-integrated MN is querying DHCP using as key 
the realm of NAI (i.e. example.com).  I think this assumption is 
correct, no?

If yes, how does the responder (DHCP, AAA) know that in example.com
there are several subnets and that each has a different prefix and a
different set of HAs on each link?

Alex

_______________________________________________
Mip6 mailing list
Mip6@ietf.org
https://www1.ietf.org/mailman/listinfo/mip6