Re: [netext] Consensus call: Work on specifying prefix delegation for Proxy Mobile IPv6?

[Alexandru Petrescu <[e-mail]>]

Le 18/08/2011 18:56, Sri Gundavelli a écrit :
>> #2 Assigning MNP to NEMO Mobile Router = RFC3963. NEMO MR per
>> definition is CMIP enabled.
>
> To ensure the terminology is right:
>
> Delegated Prefix - Prefixes hosted by the mobile node, or the network
> elements behind the mobile node.

Ok.  Sounds as the MNP to me ("Mobile Network Prefix"), sent by MR in RA
on its ingress interface, and used by LFN to form address.  Or maybe
stored by MR in its dhcpd.conf, runs as DHCP Server, and LFN as Client.

> Hosted Prefixes - prefixes hosted by the PMIPv6 mobility elements on
> the MN-AR access link. These are not delegated prefixes. An IP host
> behind the mobile node cannot use this prefix to generate an
> address, it wont receive RA's with these PIO's.

Hosted Prefix sounds as the HNP to me ("Home Network Prefix", or so,
typical of PMIP).

> HNP typically implied prefixes delivered on PMIPv6 signaling plane.

Yes, except that PMIP does not specify how that HNP is assigned, and it
could be done with DHCP.  For example, LMA could do DHCP-PD to obtain
that HNP, and put that HNP in the PBU/PBA; but still use that DHCP-PD to
obtain the prefix.  I believe currently the PMIP may need something like
this.

> If DHCP PD is used by MN or a node behind for obtaining prefixes,
> those are simple IP prefixes. However, if mobility is provided to
> those prefixes, in the form of this draft, we can group them as
> HNP's, as mobility is provided and those prefixes are anchored on
> the LMA, from routing perspective.

YEs, it could be so.  But I am not sure the best way is to anchor all
prefixes at LMA.  We could give freedom to have some prefixes "anchored"
at another entity than the LMA, maybe the DHCP Server, or maybe not even
the DHCP Server but some Gateway.

> MN/MR Distinction is clear I assume. But, NEMO MR, I may have
> implied, as mobile router with CMIP functionality in my prev mail.
> But, probably NEMO is a generic term. Any case, the distinction is
> understood, with or without CMIP ...

Yes, NEMO is a too generic term, I confuse it some times too.  But sure
CMIP/NEMO does not assign prefixes.  (there is recently a different RFC
doing DHCP-PD for NEMO, but this PD is not in the base spec NEMO RFC).

Alex

>
> Sri
>
>
>
>
>
>
>
> On 8/18/11 9:14 AM, "Sri Gundavelli"<[e-mail]>  wrote:
>
>> Alex:
>>
>> If I may comment.
>>
>>
>>> Please specify whether this prefix delegation feature is for the
>>> goal of
>> supporting Network Mobility with PMIP?
>>
>> #1 Implies, mobility for the delegated prefixes
>>
>>> Or is it to assign the HNP to the Mobile Host (not necessarily
>>> to assign
>> MNP for NEMO Mobile Router)?  The two goals are distinctive IMHO.
>>
>> Assigning HNP to mobile = mobility + delegated prefix (Same as #1)
>>
>> #2 Assigning MNP to NEMO Mobile Router = RFC3963. NEMO MR per
>> definition is CMIP enabled.
>>
>>
>> So, the draft is supporting #1.
>>
>>
>> Sri
>>
>>
>>
>>
>> On 8/18/11 8:50 AM, "Alexandru
>> Petrescu"<[e-mail]> wrote:
>>
>>> Hello Raj,
>>>
>>> Le 10/08/2011 23:34, [e-mail] a écrit :
>>>>
>>>> At IETF81, Carl Williams presented the I-D: "Prefix Delegation
>>>> for Proxy Mobile IPv6"<draft-zhou-netext-pd-pmip-01.txt>
>>>>
>>>> General consensus at the Netext WG meeting was that prefix
>>>> delegation is a required feature for PMIP6.
>>>
>>> Please specify whether this prefix delegation feature is for the
>>> goal of supporting Network Mobility with PMIP?
>>>
>>> Or is it to assign the HNP to the Mobile Host (not necessarily
>>> to assign MNP for NEMO Mobile Router)?  The two goals are
>>> distinctive IMHO.
>>>
>>> This to help formulate a problem for prefix delegation for PMIP.
>>>
>>> [...]
>>>> We are now following up with the questions on the ML.
>>>>
>>>> Question to WG:
>>>>
>>>> 1. Should the WG specify prefix-delegation support for PMIP6?
>>>>
>>>> Yes   [ ] No    [ ]
>>>
>>> Yes, if it is for MNP for Mobile Router.
>>>
>>>> 2. Can we adopt as WG document the solution proposed in I-D:
>>>> draft-zhou-netext-pd-pmip-01.txt as the starting point of this
>>>>  feature?
>>>>
>>>> Yes   [ ] No    [ ]
>>>
>>> No, unless the problem is clearer.
>>>
>>> I hope this helps.
>>>
>>> Alex
>>>
>>>>
>>>> Please respond by August 18th on the ML.
>>>>
>>>> -Chairs
>>>>
>>>> Please see the discussion at the IETF81 WG meeting on this
>>>> topic at:
>>>> http://www.ietf.org/proceedings/81/minutes/netext.txt
>>>>
>>>>
>>>> _______________________________________________ netext mailing
>>>> list [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/netext
>>>>
>>>
>>> _______________________________________________ netext mailing
>>> list [e-mail]
>>> https://www.ietf.org/mailman/listinfo/netext
>>
>> _______________________________________________ netext mailing list
>> [e-mail] https://www.ietf.org/mailman/listinfo/netext
>
>