draftietftcpmalternativebackoffecn06 Naeem Khademi Michael Welzl Grenville Armitage Gorry Fairhurst TCPM IETF 101 19 03 2018 1 Thank you for comments We had a detailed review from Michael ID: 798137
Download The PPT/PDF document "TCP Alternative Backoff with ECN (ABE)" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
TCP Alternative Backoff with ECN (ABE)draft-ietf-tcpm-alternativebackoff-ecn-06Naeem Khademi, Michael Welzl, Grenville Armitage, Gorry Fairhurst
TCPM @ IETF 10119. 03. 2018
1
Slide2Thank you for comments!We had a detailed review from Michael Scharf before WGLC and updated the draftDuring WGLC
, from Richard Scheffenegger:Some of the I-D references are already RFCs (also M. Tüxen)"I'm wondering if some generic rules-of-thumb, as to what a reasonable beta_loss vs. beta_ecn adjustment would be in this RFC might be in order (although I agree, that CCs should come up with reasonable guidance there)."
Our answer: it really depends on the CC
Note: our draft already says "The results of these tests indicate that CUBIC connections benefit from beta_{ecn} of 0.85
(cf. beta_{loss} = 0.7)"
2
Slide3Comments from Markku KojoWrong statement in section 4.1 ("Why Use ECN to Vary the Degree of Backoff?") related to timeoutWe'll remove this paragraphSpecify what happens when cwnd == ssthreshSuggest to be conservative + conform with previous versions: Congestion Avoidance only, which is only clearly the case when cwnd > ssthresh
Explain that there is a "grey area" that, in RFC 5681 style, "may benefit from additional attention, experimentation and specification."Suggest to include cwnd
<=
ssthresh in this
3
Slide4Comment from Markku Kojo /2Concern about lower bound of 2*SMSSWe will clarify that our modified backoff factor applies to adjusting ssthresh and cwnd upon receipt of ECN markAs before, cwnd may be reduced below ssthreshABE is only about changing the backoff factor
4