[dhcwg] bcmc status and issues

As was discussed in the wg meeting, a lot has changed in the bcmc
draft since it passed wglc, and there are several outstanding
issues. Here is my take on the status and the remaining issues.

The two drafts draft-ietf-dhc-bcmcv4-option-00.txt and
draft-ietf-dhc-bcmcv6-option-00.txt passed wglc in February.

They were replaced by a common draft
draft-ietf-dhc-bcmc-options-00.txt at end of March. Later in
May, June and August we got rev 01, 02 and 03 resp. See
http://tools.ietf.org/wg/dhc/draft-ietf-dhc-bcmc-options/ for
diffs.

draft-ietf-dhc-bcmc-options-01.txt through IETF LC in June.

Thomas Narten posted his review comments of 01 to DHC list
June 16th. Version 02 was made to address those.

Elwyn Davies reviewed 01 and 02 June/July and had a number
of issues. I'm sending that in a separate mail. There is
also an additional mail I will forward from Elwyn after
talking to author this last IETF.

I'm also posting a mail from Kuntal Chowdhury where he
lists the issues and resolutions in 03.

Below are the main outstanding issues / discussion points I'm
aware of. See also the i-d tracker. In addition Ralph and I
have some minor issues (I think just typos and minor textual
improvements), but the below issues are what I think we need
to discuss at this point.

I1 FQDNs

   Should FQDNs be allowed and is it ok to have multiple? If
   FQDNs are allowed, is it okay using two different option
   codes as proposed?

I2 Clarifications regarding MAYs/SHOULDs

   From mail from Brian Carpenter:

> The -03 draft says:
>
>   The option MAY contain multiple domain names, but these domain names
>   SHOULD be used to construct SRV lookups as specified in [BCMCS],
>   rather than querying for different A records.  The client can try any
>   or ALL of the domain names to construct the SRV lookups.  The list of
>   domain names MAY conatin the domain name of the access provider and
>   it's partner networks that also offer broadcast and multicast
>   service.
>
> This is not algorithmic. It does not say when it is appropriate to exercise
> either of the MAYs and does not say when it is justified to ignore
> the SHOULD. But an implementor needs these explanations to know what to
> code or configure. This is especially important for the SHOULD - do I write
> code to try all of the names, all but three, or only the fifth one?

Stig

_______________________________________________
dhcwg mailing list
dhcwg@ietf.org
https://www1.ietf.org/mailman/listinfo/dhcwg