Re: Last Call: <draft-ietf-6man-rfc4291bis-07.txt> (IP Version 6 Addressing Architecture) to Internet Standard

[David Farmer <farmer@xxxxxxx>]

On Tue, Feb 14, 2017 at 7:31 PM, Lorenzo Colitti <lorenzo@xxxxxxxxxx> wrote:

On Tue, Feb 14, 2017 at 11:18 AM, David Farmer <farmer@xxxxxxx> wrote:

However, the Interface ID of all unicast addresses is required to be 64 bit with the exception of the following; addresses for point-to-point links [RFC6164], Network-Specific Prefixes used for IPv4/IPv6 Translators [RFC6052], and addresses that start with the binary value 000.

I don't think it makes sense to cite RFC6052 here, since in RFC 6052 addresses it's not really possible to define the IID in way that makes sense. See https://mailarchive.ietf.org/arch/msg/ipv6/lyZl3I4rXhnmXYFCAQTRWyohfyI .

I conceded that is probably true for most of the IPv4-Embedded IPv6 Address Formats in section 2.2 of RFC6052, but the /96 format seems indistinguishable from other IPv6 addresses with embedded IPv4 addresses described in Section 2.4.5 of this draft.

And the following text from from section 2.4.4 of this draft seems to strongly imply that IPv6 address with embedded IPv4 addresses have an IID length other than 64.

   As noted in Section 2.4, all Global Unicast addresses other than

   those that start with binary 000 have a 64-bit interface ID field

   (i.e., n + m = 64), formatted as described in Section 2.4.1.  Global

   Unicast addresses that start with binary 000 have no such constraint

   on the size or structure of the interface ID field.

   Examples of Global Unicast addresses that start with binary 000 are

   the IPv6 address with embedded IPv4 addresses described in

   Section 2.4.5.....

Further, the fact that RFC6052 describes it as a /96 prefix, also seems to strongly imply a 32 bit IID length.  Now I can imagine counter arguments along the line that IPv6 addresses with embedded IPv4 addresses don't have IIDs either, but pulling on that string puts me at a loss to explain the rationale of the exclusion for addresses that start with binary 000 at all.  

So, I suppose we limit the statement to the /96 Network-Specific Prefixes.

Also, as I've read through RFC6052 several times, I wonder if there should be a reference to RFC6052 added in section 2.4.5. of this draft.  Maybe something like the follow added to the end of the paragraph for section 2.4.5 .

Additional IPv6 address that carry an IPv4 address are defined in IPv6 Addressing of IPv4/IPv6 Translators [RFC6052].

Thanks.

--

===============================================
David Farmer               Email:farmer@xxxxxxx
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
===============================================