Re: [v6ops] address representations, and addresses without IIDs

[Alexandre Petrescu <[e-mail]>]

Le 24/03/2021 à 15:36, Gert Doering a écrit :
> HI,
> 
> On Wed, Mar 24, 2021 at 03:32:08PM +0100, Alexandre Petrescu wrote:
>> Le 24/03/2021 à 13:11, Gabor LENCSE a écrit :
>>> I meant that you need to do the address synthesis manually. I
>>> intended the 64:ff9b::/96 WKP only as an example.
>>
>> As a side note,  I think that 64:ff9b::/96 prefix might need a 32bit
>> IID, which is probably forbidden by the IPv6 Addressing Architecture RFC
>> 4291 ("For all unicast addresses, except those that start with the
>> binary value 000, Interface IDs are required to be 64 bits long").
> 
> 64:: does start with 000

ah sorry, I missed again something in my calculator - that's my 
expertise :-)

0000 0000 0110 0100 is indeed 64::

But it is not my fault.

I have a big problem in talking about addresses that start with 000 
binary because there are so many ways to express these addresses in an 
IPv6 literal format.  Also the 1000::/3 start with 000, also 1FFF::/3 
start with 000, also 64::/96 start with 000.

Because of that problem of talking about addresses, and because of the 
000 exception, it is difficult to talk about IPv6 64bit boundaries.

This problem is not apparent only when talking about 64bit boundaries, 
but also when talking about ULAs FC an FD.

> And it has no IID at all, not "/32".

An address with no IID at all?

Bon, ok...

Alex

> 
> Gert Doering
>          -- NetMaster
>