Last Call Review of draft-ietf-bmwg-dcbench-methodology-07
review-ietf-bmwg-dcbench-methodology-07-genart-lc-romascanu-2017-06-08-00

I am the assigned Gen-ART reviewer for this draft. The General Area
Review Team (Gen-ART) reviews all IETF documents being processed
by the IESG for the IETF Chair.  Please treat these comments just
like any other last call comments.

For more information, please see the FAQ at

<https://trac.ietf.org/trac/gen/wiki/GenArtfaq>.

Document: draft-ietf-bmwg-dcbench-methodology-??
Reviewer: Dan Romascanu
Review Date: 2017-06-08
IETF LC End Date: 2017-06-13
IESG Telechat date: Not scheduled for a telechat

Summary:

This Informational I-D describes test and evaluation methodology and
measurement techniques for physical network equipment in the data center. It's
an useful and clearly written document and it is ready for publication, after a
number of minor issues and nits will be addressed.

Major issues:

Minor issues:

1. There is one single mention of the scope being benchmarking of 'physical
network equipment in the data center'. I assume this is written as opposed to
virtual networking equipment. This is not included in the title or repeated any
place else. I suggest to make more clear the scope by mentioning 'physical
equipment' at least once more, for example in the Introduction section.

2. Section 1.2: I am a little concerned about adding additional interpretations
to the keywords in RFC2119.

Especially about:

'MAY: Comprehensive metric for the scenario described'

MAY has a different meaning in 2119, and the fact that a metric is
'comprehensive' (whatever this means) seems to me orthogonal to the 2119
semantics. The proposed interpretation makes little sense to me, if
'comprehensive' than why not required? or at least as a 'should'?

3. Section 6.2: 'The intensity of a microburst MAY be varied ...' It would be
useful to make clear what is meant by 'intensity'.

Nits/editorial comments:

1. The different formatting of the references [1], [2], etc. vs. [RFC2119],
etc. is slightly confusing. Unless there is some good reason I suggest to fix
this. Also use of references is inconsistent through the text, sometimes they
are mentioned, sometimes they are not.

2. Section 2.2: 'Alternatively when a traffic generator CAN NOT be connected '
- this capitalization is not conformant to RFC2119.

3. 'vlan ZZZ' is not consistent with 'vlan X' and 'vlan Y' previously used.

4. Missing reference - RFC 6985

5. Missing expansion and references for DSCP and COS

6. Sections 3.3 and 6.3: s/number of iteration/number of iterations/

7. The methodology (e.g. in section 5) makes an assumption that the DUT has at
least nine ports. This may be trivial in a DC, but it's worth being mentioned