[{"author": "Jeffrey Haas", "text": "<p>I would like to wander over from lsr for the presentations on the yang modules.  Could someone give 2 min warning before those presentations?</p>", "time": "2022-07-29T14:10:43Z"}, {"author": "Yoshifumi Nishida", "text": "<p>ok. I will send message before the presentation</p>", "time": "2022-07-29T14:12:33Z"}, {"author": "Jeffrey Haas", "text": "<p>Thank you.</p>", "time": "2022-07-29T14:13:01Z"}, {"author": "Neal Cardwell", "text": "<p>Clarification: the \"issue 3\" is fixed in my draft patches, but not in Linux yet. The setting of alpha_cubic to 1 after cwnd &gt; w_max or prior_cwnd is not in Linux yet.</p>", "time": "2022-07-29T14:19:33Z"}, {"author": "Rodney Grimes", "text": "<p>Can you run Hystart++ with Cubic?</p>", "time": "2022-07-29T14:23:18Z"}, {"author": "Yi Huang", "text": "<p>beta_cubic in windows TCP is 0.7 as well. JFYI.</p>", "time": "2022-07-29T14:23:25Z"}, {"author": "Jonathan Morton", "text": "<p>@Rodney - I don't see why not</p>", "time": "2022-07-29T14:23:38Z"}, {"author": "Rodney Grimes", "text": "<p>Does that solve the Alpha issue of issue 1?</p>", "time": "2022-07-29T14:23:56Z"}, {"author": "Jonathan Morton", "text": "<p>it would be more related to issue 2 than 1</p>", "time": "2022-07-29T14:24:28Z"}, {"author": "Jonathan Morton", "text": "<p>issue 1 is about the Reno compatibility mode</p>", "time": "2022-07-29T14:24:46Z"}, {"author": "Rodney Grimes", "text": "<p>I miss typed, I  meant issue 2</p>", "time": "2022-07-29T14:25:19Z"}, {"author": "Vidhi Goel", "text": "<p>@Rodney, yes. And this is mentioned in the draft as well</p>", "time": "2022-07-29T14:27:28Z"}, {"author": "Rodney Grimes", "text": "<p>If there are these issues perhaps instead of going to PS, why not take 8312bis to EXP, and get this data, then go to PS.</p>", "time": "2022-07-29T14:27:57Z"}, {"author": "Jonathan Morton", "text": "<p>Issue 1:  the current value of C in CUBIC_alpha is widely deployed and isn't causing any obvious problems, so I don't see the need to change it for a theoretical reason</p>", "time": "2022-07-29T14:29:59Z"}, {"author": "Jonathan Morton", "text": "<p>Issue 2: likewise 0.7 is widely deployed and isn't causing any practical problems, since in a deep buffer HyStart is likely to trigger before the buffer is full</p>", "time": "2022-07-29T14:30:45Z"}, {"author": "Jonathan Morton", "text": "<p>while a shallow buffer is likely to be associated with AQM which likely triggers before the buffer is full</p>", "time": "2022-07-29T14:31:20Z"}, {"author": "Rodney Grimes", "text": "<p>@Jonathan, so that would explain WHY we dont see an issue with the 0.7, and in effect that value just wont matter due to HyStart kicking in?</p>", "time": "2022-07-29T14:31:40Z"}, {"author": "Jonathan Morton", "text": "<p>correct</p>", "time": "2022-07-29T14:31:52Z"}, {"author": "Jonathan Morton", "text": "<p>so CUBIC_beta should be tuned for congestion avoidance mode rather than slow-start exit</p>", "time": "2022-07-29T14:32:18Z"}, {"author": "Rodney Grimes", "text": "<p>So the problem is \"Hidden\".  <em>sigh</em></p>", "time": "2022-07-29T14:33:06Z"}, {"author": "Jonathan Morton", "text": "<p>remind me to discuss this with you elsewhere</p>", "time": "2022-07-29T14:35:14Z"}, {"author": "Neal Cardwell", "text": "<p>Thanks for the clarification Bob. Can you please point to the slide numbers in your ICCRG slides that have Reno and CUBIC in the same queue? Thanks!</p>", "time": "2022-07-29T14:41:26Z"}, {"author": "Neal Cardwell", "text": "<p>AFAICT those are all with an AQM, and I think Markku's concern with alpha mainly pertained to CUBIC and Reno coexisting in that  drop-tail FIFO.</p>", "time": "2022-07-29T14:42:35Z"}, {"author": "Neal Cardwell", "text": "<p>Note in Markku's email <a href=\"https://mailarchive.ietf.org/arch/msg/tcpm/pN1fARPVzqcDlt--xiMoDm_loAM/\">https://mailarchive.ietf.org/arch/msg/tcpm/pN1fARPVzqcDlt--xiMoDm_loAM/</a> the phrase: \" he model is incorrect for the traditional and still today prevailing tail-drop router case\"</p>", "time": "2022-07-29T14:45:33Z"}, {"author": "Vidhi Goel", "text": "<p>I believe its slide 8 in Bob's slides in ICCRG.</p>", "time": "2022-07-29T14:47:43Z"}, {"author": "Neal Cardwell", "text": "<p>If I'm reading that correctly, that's showing \"ECN-CUBIC\" vs Reno, which again sounds like an AQM (to get the ECN marking) rather than drop-tail.</p>", "time": "2022-07-29T14:54:04Z"}, {"author": "Gorry Fairhurst", "text": "<p>I left the queue</p>", "time": "2022-07-29T14:55:40Z"}, {"author": "Vidhi Goel", "text": "<p>there is ECN-CUBIC and then there is just CUBIC. I am not sure if just CUBIC uses any AQM</p>", "time": "2022-07-29T14:55:54Z"}, {"author": "Jonathan Morton", "text": "<p>I don't think any of the graphs shown in the Prague CC material involve a drop-tail bottleneck</p>", "time": "2022-07-29T14:56:48Z"}, {"author": "Neal Cardwell", "text": "<p>The \"Reno\" in slide 8 is only on the second row, which only has \"ECN-CUBIC\" results, not \"CUBIC\" results.</p>", "time": "2022-07-29T14:57:50Z"}, {"author": "Neal Cardwell", "text": "<p>My sense agrees with Jonathan's, that none of the graphs shown in the Prague CC material involve a drop-tail bottleneck.</p>", "time": "2022-07-29T14:58:15Z"}, {"author": "Gorry Fairhurst", "text": "<p>ACK filtering change worked for me, thanks.</p>", "time": "2022-07-29T14:59:47Z"}, {"author": "Yoshifumi Nishida", "text": "<p>accn presentation is toward the end</p>", "time": "2022-07-29T15:01:05Z"}, {"author": "Yoshifumi Nishida", "text": "<p>next wiil be yang</p>", "time": "2022-07-29T15:01:46Z"}, {"author": "Neal Cardwell", "text": "<p>Clarification, this Linux code we tested this week was Ilpo's AccECN code.</p>", "time": "2022-07-29T15:05:07Z"}, {"author": "Chris Box", "text": "<p>Yang presentation is starting</p>", "time": "2022-07-29T15:19:06Z"}, {"author": "Lars Eggert", "text": "<p>keepalives?</p>", "time": "2022-07-29T15:35:08Z"}, {"author": "Rodney Grimes", "text": "<p>Keepalive and dead timer should tear down the BGP sessession.</p>", "time": "2022-07-29T15:35:58Z"}, {"author": "Michael T\u00fcxen", "text": "<p>Application layer keep alives / heartbeat</p>", "time": "2022-07-29T15:35:58Z"}, {"author": "Wesley Eddy", "text": "<p>or ... more YANG</p>", "time": "2022-07-29T15:36:08Z"}, {"author": "Rodney Grimes", "text": "<p>This \"case #1\" was solved over 30 years ago.</p>", "time": "2022-07-29T15:36:44Z"}, {"author": "Lars Eggert", "text": "<p>i don't understand why a broken application needs to be fixed by having another application look at arbitrary TCP state</p>", "time": "2022-07-29T15:42:13Z"}, {"author": "Jeffrey Haas", "text": "<p><span class=\"user-mention\" data-user-id=\"35\">@Lars Eggert</span>  The application itself is often not broken.  It's broken TCP and the application is the victim.</p>", "time": "2022-07-29T15:51:02Z"}, {"author": "Jeffrey Haas", "text": "<p>And while we have the headaches that come on occasion from modifying tcp stacks, we also have similar issues with bugs in stock Linux stacks.</p>", "time": "2022-07-29T15:52:01Z"}, {"author": "Jeffrey Haas", "text": "<p>You develop a different perspective on the bugs when your tcp sessions are often up for months.</p>", "time": "2022-07-29T15:52:15Z"}, {"author": "Jonathan Morton", "text": "<p>in the case mentioned, the application was hung but the TCP stack was still operating</p>", "time": "2022-07-29T15:52:17Z"}, {"author": "Jeffrey Haas", "text": "<p>Understood that was the example.  I'm offering another example.</p>", "time": "2022-07-29T15:52:35Z"}, {"author": "Jeffrey Haas", "text": "<p>The hung state from a wedged receiver has a discussion for a \"send holdtimer\" to cover this condition in the IDR mailing list.</p>", "time": "2022-07-29T15:54:01Z"}, {"author": "Gorry Fairhurst", "text": "<p>Why do we need R so large? .. just because it is possible???? .. ona 1 Gbps link, we can have 1 ACK every 100 or so packets, with little overhead=</p>", "time": "2022-07-29T15:54:39Z"}, {"author": "Gorry Fairhurst", "text": "<p>In another case: A large value is harmful also when there is path flapping or a path change.</p>", "time": "2022-07-29T16:01:57Z"}, {"author": "Neal Cardwell", "text": "<p>80Gbps * 100ms is a BDP of 660,502 packets, at MTU=1500bytes. In which case an ACK per 2,000 packets would give 330 ACKs per round trip. So IMHO a max value of 2047 is reasonable.</p>", "time": "2022-07-29T16:01:57Z"}, {"author": "Gorry Fairhurst", "text": "<p>Is it reasonable from a CC perspective?</p>", "time": "2022-07-29T16:02:28Z"}, {"author": "Jonathan Morton", "text": "<p>yes, this was the thrust of the mathematics I posted on the list</p>", "time": "2022-07-29T16:02:29Z"}, {"author": "Neal Cardwell", "text": "<p>Yes, it's reasonable from a CC perspective if the CC is using pacing, which it ought to be.</p>", "time": "2022-07-29T16:02:51Z"}, {"author": "Neal Cardwell", "text": "<p>As Jonathan noted, if there is loss, the receiver will ACK immediately.</p>", "time": "2022-07-29T16:03:05Z"}, {"author": "Jonathan Morton", "text": "<p>if congestion feedback is generated by the receiver or on the path, this will result in immediate acks regardless of the value of R</p>", "time": "2022-07-29T16:03:07Z"}]