CUBIC for Fast Long-Distance Networks
Summary: Has enough positions to pass.
Spencer Dawkins Yes
I'm really glad to see this specification entering IESG Evaluation. I have a few comments, but most are editorial or (at most) about improving clarity. In this text, In a smaller BDP network where Standard TCP flows are working well, the absolute amount of the window decrease at a loss event is always smaller because of the multiplicative decrease. I got lost on "always smaller" than what - than Standard TCP? Or CUBIC? Or something else? Nit: Is "weithed" "weighted", or is it something else? In 4.7. Timeout In case of timeout, CUBIC follows the standard TCP to reduce cwnd, but sets ssthresh using beta_cubic (same as in Section 4.5). should "standard TCP" be "Standard TCP"? I wasn't watching for other occurrences of "standard TCP", but I noticed a bunch of them. In this text, CUBIC MUST employ a slow start algorithm, when the cwnd is no more than ssthresh. Among the slow start algorithms, CUBIC MAY choose the standard TCP slow start [RFC5681] in general networks, or the limited slow start [RFC3742] or hybrid slow start [HR08] for high-bandwidth and long-distance networks. is there any guidance you can give implementers about when to choose specific slow start algorithms? In 5.10. Incremental Deployment CUBIC requires only the change of TCP senders, and does not require any assistant of routers. I'm not parsing the sentence. Is it saying CUBIC requires only changes to TCP senders, and does not require any changes to routers. ? Either way, it might be worth pointing out here that no changes to TCP receivers are required, either.
Mirja Kühlewind Yes
Deborah Brungard No Objection
Ben Campbell No Objection
I'm a confused by the fact that this draft claims to specify CUBIC, but also cites another document for CUBIC. Which is the authoritative definition? If the answer is "that other document", then some words to clarify that would be helpful. I gather CUBIC is not an acronym? If correct, why spell it in all-caps?
Benoit Claise No Objection
Some minor issues from Qin Wu's OPS DIR review needs to be taken care of (as agreed by Mirja)