Re: A proposal for draft-ietf-6man-rfc4291bis-07

[Alexandre Petrescu <[e-mail]>]

David,

Sincerely I appreciate the effort and would like to agree to most
points.  But here is my answer and I wont change.

Le 02/03/2017 à 23:43, David Farmer a écrit :
> I've been thinking about this, really hard, and listening to what
> others are saying. I think we want to say the following;
>
> 1. IPv6 unicast routing is 128 bits in length [BCP198], AKA not
> classful!

Agreed, we want to say this.

> 2. Subnet Prefixes of 64 bits are RECOMMENDED

I disagree, we dont want to say this.

> 3. IIDs are REQUIRED to be 64 bits

I disagree.

> 4. Say nothing that would imply you cannot configure a prefix length
> longer than 64 bits in length, at least manually.

I agree.

> 5. Avoid all explicit exceptions, like RFC6164, etc...

I dont know.

> How about in the current draft remove the 2nd paragraph of 2.4, then
> add the following paragraph at the end of section 2.4
>
> IPv6 unicast routing is based on prefixes of any length up to 128
> bits [BCP198]. However, subnet prefixes of 64 bits in length are
> recommended in normal use.

I agree.

> Furthermore, subnet prefixes and Interface IDs both of 64 bits in
> length are required when a unicast address is formed from a subnet
> prefix and an Interface ID.

This phrase is too much.

> Otherwise, when a subnet prefix of
> another length is used then a unicast address is assumed to have no
> internal structure and is treated as a single 128 bit quantity. The
> rationale for the 64 bit boundary in IPv6 addresses can be found in
> [RFC7421].

This is an explanation of the earlier phrase, and as such is too too much.

> Then in the 4th paragraph of 2.4.1
>
> As noted in Section 2.4, when a unicast address is formed from a
> subnet prefix and an Interface ID, both are required to be 64 bits in
> length.

Is the 'subnet prefix' coming from an RA?  From elsewhere?  There are 
very many ways to make a subnet prefix.

> I think this accomplishes the 5 goals above and is not inconsistent
> with at least the intent of RFC4291. If you read section 2.5 of the
> RFC4291, I think my new paragraph is only more clearly restating what
> is said there already.

I think your text suggestion is indeed answering the initial 
requirements formulated at the top.

But I can not agree with it as such.

Again - thank you for the effort, it is good to pursue.

Alex
>
> Note that SLACC and other things that use IIDs have 64 bits required
> for the IID and subnet prefix length.  But, it should also be noted
> that host are not required to have any knowledge of an addresses
> structure including its own addresses, therefore IIDs are in fact
> optional. Additionally, 64 bit prefix lengths are recommend for
> everything, including manual config and DHCP, but manual config and
> DHCP are allowed other prefix lengths as well.   Also, the exception
> for unicast addresses that begin with 000 is no longer really
> necessary.
>
> Rotten tomatoes and eggs now please. :)
>
> Thanks.
>
> -- =============================================== David Farmer
> Email:[e-mail] <mailto:[e-mail]> Networking &
> Telecommunication Services Office of Information Technology
> University of Minnesota 2218 University Ave SE        Phone:
> 612-626-0815 Minneapolis, MN 55414-3029   Cell: 612-812-9952
> ===============================================
>
>
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list [e-mail] Administrative
> Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
>