Updates for the Back-to-Back Frame Benchmark in RFC 2544
RFC 9004
Yes
No Objection
Note: This ballot was opened for revision 03 and is now closed.
Warren Kumari Yes
Alvaro Retana No Objection
Erik Kline No Objection
Martin Duke (was Discuss) No Objection
Thanks for addressing my DISCUSS
Murray Kucherawy No Objection
There are lots of terms throughout this document that are capitalized when I'm not certain they should be (e.g., "Buffer", "Burst", "Data Center", etc.), and I ended up looking around for specific definitions of them. You might want to give this a pass to check on all of that, because leaving it to the RFC Editor to sort out will probably slow things down.
Robert Wilton No Objection
Hi Al, Thanks for this document. I mostly found it easy to read with just a few minor comments: I agree with Murray's comment that there seems more text capitalized than is potentially helpful. Possible words/phrases to check: Obsolete, "Back-to-back Frames" except when used for "Back-to-back Frame Benchmark", "Data Center", "Throughput", "Ingress", "Egress", "Frames", "Buffer" Section 2, para 3: However, conditions simultaneously sending multiple frame sizes, such as those described in [RFC6985], MUST NOT be used in Back-to-back Frame testing. I found this sentence slightly hard to parse, rephrasing it might aid readability. In section 6, it states "The number of tests Averaged for the Benchmark, N, MUST be reported." Should that be illustrated in the "Back-to-Back Frame Results" table, either as part of the table, or as extra information alongside it? Regards, Rob
Roman Danyliw No Objection
Thank you for discussing the feedback from the SECDIR review and thank you to Mališa Vučinić for this review. The candidate text proposed at https://mailarchive.ietf.org/arch/msg/secdir/2CTnqKT7kwGLLSKMnEUTP5HC8wc/ is helpful (fully recognizing that this is addressing a issue not specified to this particular benchmark update).
Éric Vyncke No Objection
(Barry Leiba; former steering group member) No Objection
(Benjamin Kaduk; former steering group member) No Objection
Thank you for staging changes to resolve the tsvart and secdir reviewers' comments. I assume that it goes without saying that the computed StdDev are sample standard deviations, not population standard deviations. Section 6 What kind of DUT is only going to have 4 microseconds of buffer? Section 10.1 It is surprising to see a normative reference being made to the Obsolete RFC 1944. Likewise, some of the references are only tenuously referenced and would seem better characterized as informative (e.g., RFC 6201, RFC 6985). Section 10.2 On the other hand, if the methods of RFC 8239 SHOULD be used in some cases, that suggests that it should more properly be characterized as a normative reference. Similarly, the TST009 binary search is one of the two options from which a selection MUST be made, which seems to make it normative as well.
(Deborah Brungard; former steering group member) No Objection