Re: [Cbor] I-D Action: draft-ietf-cbor-network-addresses-04.txt

Le 26/04/2021 à 07:15, Carsten Bormann a écrit :
> On 26. Apr 2021, at 02:06, Michael Richardson <mcr+ietf@sandelman.ca> wrote:
>>
>>> - so you don’t want to support the commonly used shortcut of specifying
>>> both and address and a prefix in one?
>>
>> I don't understand your suggestion.
>>     X - address
>> [n,X] - prefix
> 
> Back in IPv4-land, we often said something like
> 
> “Configure the interface to 192.168.3.5/24”
> 
> Which is a short-cut for
> 
> “Configure the interface to address 192.168.3.5, prefix 192.168.3.0/24”
> 
> Since address/prefix is structurally similar to a prefix (but uses a full 32-bit address), it could be expressed by adding a third case
> 
>>     A - address
>> [n,P] - prefix
>> [A,n] - address/prefix
> 
> Where A is a 4-byte byte-string, P is a 0-to-4-byte byte string, and n is an unsigned integer 0 to 32.
> 
> This construct is less common in IPv6-land, because n is usually 64,

well, no, but I think it was said.

> so sending the address suffices, but there are exceptions to that here, too.

YEs, there are exceptions.  There is a VSLAAC I-D (V for Variable 
length) and open source linux kernel implementation which configures an 
IP address out of a prefix of any length, not only 64.  Manually it is 
also possible.  OpenBSD also has such code.

Alex
> 
> Grüße, Carsten
> 
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> ipv6@ietf.org
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
> 