Re: [v6ops] RFC7084

[Alexandru Petrescu <[e-mail]>]

Le 11/12/2013 19:46, Owen DeLong a écrit :
>
> On Dec 11, 2013, at 1:45 AM, Alexandru Petrescu <[e-mail]> wrote:
>
>> Le 10/12/2013 18:44, Ole Troan a écrit :
>>>>> In no case do I believe that M or O provide any indication about
>>>>> IA_PD.
>>>>
>>>> You should read RFC 7084 again, then.
>>>>
>>>> Standards say what they say, not what you think they should say!
>>>> :)
>>>
>>> RFC7084 isn't a standard. can we please stop this now. having the M/O
>>> debate one more time is unlikely to provide a different result than
>>> the previous times we have had this debate.
>>
>> I agree we should go past the typical M/O debate, learn from it.
>>
>> One way I read the ongoing discussion, if I am not wrong, is that there
>> is difficulty created by the lack of a flag specific to IA_PD ("As far
>> as I know the M flag is linked only to IA_NA. As far as I can see IA_PD
>> is not linked to the M flag at all.”).
>
>>
>> Am I the only to read this as maybe a hint towards necessity of creation
>> of a new flag akin to M/O but specific to IA_PD? I.e. it would be allow
>> a Router to see whether it may be able to specifically request a
>> Delegated Prefix even when it would self-configure its address by other
>> means than DHCP.
>
> I don’t see the need. A router which wants a prefix should ask for one.
>
> The worst case is it doesn’t get an answer.
>
> The next worst case is it is denied.
>
> What is the benefit of having a flag to (sometimes) say that prefixes are
> available for the asking?

Just as with addresses, it may be more reliable to configure prefixes by 
starting with a beacon announcing availability.

But yes, it's just another 'what if' case.

>> (in the past some local thought was given about how RA would advertise a
>> capability of delegating prefixes, e.g. this 'D' flag in a draft:
>>
>>>     0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
>>>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>>>    |     Type      |     Code      |           Checksum            |
>>>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>>>    | Cur Hop Limit |M|O|H|Prf|P|D|r|        Router Lifetime        |
>>>    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>> )
>>
>
> I wasn’t around for that discussion, but my guess would be that the discussion
> likely came to the same conclusion I did… If you want a prefix, ask. A D flag
> wouldn’t really improve anything.

It was a local non-IETF discussion.

The later versions of that draft eliminated that D flag but I forgot 
why.  It may be as you say.

Alex

>
> Owen
>
>
>