Methodology for IP Multicast Benchmarking
draft-ietf-bmwg-mcastm-14

I did ask WG chairs for a response to the COMMENTS from OPS-DIR review (on april 16th). Sofar it seems I did not get a response. I am checking again.

Comments from the ops directorate by Pekka Savola:

==> has this document been sent for review in multicast WGs, e.g.,
MBONED?  I don't recall that.  I worry whether there are some tests
missing, or that the tests do not test the right things.

==> the methodology is not "IPv6-enabled": it discusses things like
IGMP and TTL without reference to the IPv6 counterparts.  This
should be fixed.

==> this document uses the term "frames".  Looking at the frame sizes
(up to 1518 bytes) this makes one believe we're talking about
L2 and being specific to Ethernet.  (In the IETF, we're typically
interested of packets; frames are just a thing to be dealt with :)

Wouldn't it make sense to convert to using "packets" (and adjust the
sizes appropriately), or is this in conflict with typical BMWG
methods?

4.4. Encapsulation/Decapsulation (Tunneling) Throughput

==> here, it might not hurt to spell out that this is
actually benchmarking the PIM Register -messages, and that
in and of itself the things such as encapsulation format.
(Or, are you considering DVMRP testing here?)


  4.4.3. Re-encapsulation Throughput

==> how does this relate to multicast?  AFAICS, this is only relevant
for DVMRP, which most people aren't interested in any case.

Date: Thu, 14 Aug 2003 10:10:50 -0400
From: Kevin Dubray
To: Randy Bush , Bert Wijnen
CC: Al Morton
Subject:  to AD/IESG review

ADs,

The BWMG's Methodology for IP Multicast Benchmarking I-D:

    http://www.ietf.org/internet-drafts/draft-ietf-bmwg-mcastm-13.txt

is being forwarded to you for review as a candidate for [Informational] RFC.

The 2nd editor and the task team seem to have injected a substantive amount
of quality to this work.