Re: [Autoconf] Conclusion: draft-ietf-autoconf-adhoc-addr-model-02.txt

[Thomas Heide Clausen <[e-mail]>]

Dear Alex,

On Feb 16, 2010, at 21:29 PM, Alexandru Petrescu wrote:

> Le 16/02/2010 21:23, Thomas Heide Clausen a écrit :
>> Dear Alex,
>>
>> Autoconf is about configuring addresses on interfaces, not on
>> allocating addresses in a registry, not sending IP packets to
>> multicast destinations.
>
> Thomas, thank you for the reply.
>
> Link-layer multicast mechanisms are used in any autoconfing (DHCPv6,
> SLAAC, probably more) mechanism.
>
> A stack booting up sends packets to multicast destinations.
>
> MANET already allocates a multicast address for this.
>
> Suffices it to mention it.
>

These reflections belong properly in the problem-statement/scoping and  
solution-space discussions -- hopefully, we will be able to get to  
those (the fun part: building protocols) soon. So hold that thought  
until later.

> Otherwise leave place for non-understanding: will the IPv6 autoconf
> stack use the MANET multicast address?  Or the other non-MANET  
> multicast
> address?

If I configure my addresses manually, as is one viable option, I can  
follow the recommendations in draft-ietf-autoconf-adhoc-addr- 
model-02.txt and have a valid configuration. In that case, the  
"configuration mechanism" needs no multicast. I note that this may  
apply more to IPv4 than IPv6, and that the document covers both.

If a MANET autoconfiguration protocol needs to exchange information  
for proper functioning - which it may well do - then that protocol  
will have to decide on which addresses, messages and algorithms to use  
for that. So again, these reflections belong properly in the problem- 
statement/scoping and solution-space discussions.

Sincerely,

Thomas

> I mostly agree with you.
>
> And there are two different people here (Teco, myself) saying
> approximately the same thing about multicast.  The autoconf group is  
> not
> large.  Are two opinions worth ignoring?
>
> Alex
>
>>
>> Sincerely,
>>
>> Thomas
>>
>>
>> On Feb 16, 2010, at 21:20 PM, Alexandru Petrescu wrote:
>>
>>> Le 16/02/2010 21:01, Jari Arkko a écrit :
>>>> Great. Lets move this doc forward!
>>>
>>> YEs, let's move this forward and add multicast discussion to it
>>> without which autoconf can't fly. Multicast is what typical
>>> autoconfiguration protocols use today without which they'd never
>>> work.
>>>
>>> Multicast is what IPv6 got builtin precisely for the reason of
>>> autoconfing.
>>>
>>> This draft being silent about multicast spells it's not autoconf,
>>> IMHO.
>>>
>>> Alex
>>>
>>>>
>>>> Jari
>>>>
>>>> Ryuji Wakikawa kirjoitti:
>>>>> Dear All,
>>>>>
>>>>> We have concluded the WGLC of
>>>>> draft-ietf-autoconf-adhoc-addr-model-01.txt on Dec/23/09, and
>>>>> have a -02 document issued, following up on this.
>>>>>
>>>>> Thanks to all for all the reviews and comments to this
>>>>> document!
>>>>>
>>>>> After Thomas and I (Chairs) carefully reviewed discussions on
>>>>> the mailing list, we do find that there is rough consensus for
>>>>> the current document.
>>>>>
>>>>> There was an individual objection to the description of the
>>>>> use of link-local address, but we did not detect wide support
>>>>> within the working group. This objection will, of course, be
>>>>> reflected in the PROTO write-up that will be sent to the IESG
>>>>> and the ADs, and reflected in the tracker.
>>>>>
>>>>> As a conclusion, we have established rough consensus to the
>>>>> new document.
>>>>>
>>>>> The WG chairs will start preparing the PROTO writeup for
>>>>> forwarding the document.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> WG chairs
>>>>>
>>>>>
>>>>
>>>> _______________________________________________ Autoconf mailing
>>>> list [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/autoconf
>>>>
>>>
>>> _______________________________________________ Autoconf mailing
>>> list [e-mail]
>>> https://www.ietf.org/mailman/listinfo/autoconf
>>
>>
>
> _______________________________________________
> Autoconf mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/autoconf