Re: Comments on draft-ietf-6man-multicast-addr-arch-update-02

[Stig Venaas <[e-mail]>]

Hi Jouni

On 2/10/2014 1:23 AM, Jouni Korhonen wrote:
> Sorry being a bit late but I had a read of this document.
>
> ** Some nits:
>
> s/[ADDRARCH]/[RFC4291]
>
> ** While I agree this document is needed there is a general procedural
> issue I have. This I-D patches three other RFCs, which over the time
> may become an issue. I would rather see each document having their
> own bis and RFC4291 update being the master on all of these updates. I
> recon it might be awfully late for this, though.

I agree. I'm not sure what is the right thing to do here, I've also been
wondering if what you say would be better. I've been asking this
question several time, but I haven't really got much feedback.

I wonder if it might be worth checking with the IESG or others to see
what the best way is. It would be a big task updating those documents
though. One way forward could be to publish this document soon, and then
work on bis versions of those documents that may also include other
improvements.

> ** Section 4.1:
>
>     NEW:
>
>           |   8    |  4 |  4 |   8    |    8   |       64       |    32    |
>           +--------+----+----+--------+--------+----------------+----------+
>           |11111111|flgs|scop|reserved|  plen  | network prefix | group ID |
>           +--------+----+----+--------+--------+----------------+----------+
>                               ^^^^^^^^
>
> Shouldn't this also be align with other NEW formats? Others have a new set
> of "flgs" + "rsvd" here.

You're right, will update this.

> ** In all proposed updated address formats we seem to have the following:
>
>      |   8    |  4 |  4 |  4 |..
>      +--------+----+----+----+----+--------+----------------+----------+
>      |11111111|flgs|scop|flgs|..                                       |
>      +--------+----+----+----+----+--------+----------------+----------+
>                ^^^^^^^^^^^^^^
>
> I find this confusing having two distinct sets of "flgs". Why not concatenate
> that to one 12 bits field? Or at least name the latter "flgs" differently?

I'm thinking of maybe calling them flag field 1 and flag field 2 and
use the names ff1 and ff2 in the diagram. Thoughts on that?

> Either way current statements like in Section 4.2
>
>                                           +-+-+-+-+
>           flgs is a set of four flags:    |X|R|P|T|
>                                           +-+-+-+-+
>
> are wrong, since there are now two sets of "flgs" that sum up to eight flags.

So maybe say that flag field one contain 4 flags...

>
> Therefore I would propose for example the following format:
>
>      |   8    |      12      | ..
>      +--------+----+----+----+----+--------+----------------+----------+
>      |11111111|flgs_and_scop | ..                                      |
>      +--------+----+----+----+----+--------+----------------+----------+
>
> Then describe the "flgs_and_scop" as
>
>      11       7       3     0
>      +-+-+-+-+-+-+-+-+-+-+-+-+
>      |X|R|P|T| scope |r|r|r|r|
>      +-+-+-+-+-+-+-+-+-+-+-+-+
>
> where "rrrr" are for future assignment as additional flag bits.
>
> Or alternatively:
>
>      |   8    |  4 |  4 |  4 |..
>      +--------+----+----+----+----+--------+----------------+----------+
>      |11111111|flgs|scop|newf|..                                       |
>      +--------+----+----+----+----+--------+----------------+----------+
>
> where "newf" are for future assignment as additional flag bits.

I prefer this latter one. What do you think of flag field 1 and
flag field 2 versus your proposal?
>
>
> ** Section 5 IANA Considerations:
>
>     This document may require IANA updates.  However, at this point it is
>     not clear exactly what these updates may be.
>
> I find this statement rather sloppy. At least the authors should
> give a try listing possible impacts & updates.

Oops. This was true at some point and I guess we overlooked this later.
I agree, it is pretty sloppy. I need to think a bit about this,

Thanks for your comments. I intend to submit a new version before
Friday.

Stig

> - Jouni
>