Re: Objection to draft-ietf-6man-rfc4291bis-07.txt

[Brian E Carpenter <[e-mail]>]

On 02/03/2017 06:03, Iván Arce wrote:
> El 1/3/17 a las 6:51, Lorenzo Colitti escribió:
> 
>>
>>         Another thing I think we should avoid is to remove the fixed 64
>>         barrier and open the door to having this debate again and again,
>>         once for every new IPv6-over-foo document and once for every new
>>         address configuration protocol (today we have SLAAC and DHCPv6,
>>         who knows what we'll have in the future).
>>
>>
>>     Which is why it time to get this right and saying it is now and
>>     forever 64 isn't right.
>>
>>
>> Do you agree that a fixed boundary is useful or not? For 20 years the
>> standards have guaranteed that 64 bits of IIDs were available to hosts
>> that wanted to use them. If we make that barrier mobile, there will be
>> no guarantee in the standards any more. Who should be allowed to set the
>> boundary? An IPv6-over-foo document? An address configuration technology
>> such as SLAAC? 
> 
> The last paragraph of section "2.5.1 Interface identifiers" in 4291 said:
> 
>    The details of forming interface identifiers are defined in the
>    appropriate "IPv6 over <link>" specification, such as "IPv6 over
>    Ethernet" [ETHER], and "IPv6 over FDDI" [FDDI].
> 
> The corresponding paragraph in rfc4291bis is:
> 
>    The details of forming interface identifiers are defined in other
>    specifications, such as "Privacy Extensions for Stateless Address
>    Autoconfiguration in IPv6" [RFC4941] or "A Method for Generating
>    Semantically Opaque Interface Identifiers with IPv6 Stateless Address
>    Autoconfiguration (SLAAC)"[RFC7217].  Specific cases are described in
>    appropriate "IPv6 over <link>" specifications, such as "IPv6 over
>    Ethernet" [RFC2464] and "Transmission of IPv6 Packets over ITU-T
>    G.9959 Networks" [RFC7428].  The security and privacy considerations
>    for IPv6 address generation is described in [RFC7721].
> 
> Don't you agree with that?

Iván,

That is the basis for Alexandre's suggestion to simply drop the mention
of 64 from the architecture: because it's an implementation parameter.
With my Computer Scientist hat on, I think that's the right thing to do.
With my Pragmatist hat on, I think stating that /64 is recommended is
a more practical approach. Stating that it's required is plain wrong.

    Brian