[{"author": "Spencer Dawkins", "text": "<p>Multi-hop is ALWAYS an issue ... even in human space.</p>", "time": "2025-07-21T12:39:37Z"}, {"author": "Christian Huitema", "text": "<p>@lars, if you want to run real code, you may want to try <a href=\"https://github.com/private-octopus/picoquic_ns\">https://github.com/private-octopus/picoquic_ns</a></p>", "time": "2025-07-21T12:49:15Z"}, {"author": "Mohit Tahiliani", "text": "<p>We had a discussion with one of the teams during hackathon to explore the possibility of linking picoquic with ns-3. ns-3 does not have a native model of QUIC</p>", "time": "2025-07-21T12:54:46Z"}, {"author": "Christian Huitema", "text": "<p>@mohit -- I would be happy to support people doing with that.</p>", "time": "2025-07-21T12:56:47Z"}, {"author": "Mohit Tahiliani", "text": "<p>Thanks, Christian. I will get in touch with you. It will be our team of students.</p>", "time": "2025-07-21T12:58:56Z"}, {"author": "Joseph Beshay", "text": "<p>One side note is that the ProbeRTT interval does not impact the MinRTT Filter Length which is currently 10s. The 10s probing interval reduces the frequency of probing but the impact on the min rtt value of the model is arguable smaller.</p>", "time": "2025-07-21T13:15:30Z"}, {"author": "Tim Chown", "text": "<p>what's the status of BBRv3 in the Linux kernel?</p>", "time": "2025-07-21T13:18:24Z"}, {"author": "Christian Huitema", "text": "<p>Probe RTT is not just about the measure of min RTT by the local implementation. It has the side effect of enabling <em>other</em> connections to find the \"real\" min RTT.</p>", "time": "2025-07-21T13:20:12Z"}, {"author": "Neal Cardwell", "text": "<p>Hi, Tim. Re \"what's the status of BBRv3 in the Linux kernel?\"; it is not in mainline Linux; it's in github: <a href=\"https://github.com/google/bbr/blob/v3/README.md\">https://github.com/google/bbr/blob/v3/README.md</a></p>", "time": "2025-07-21T13:22:22Z"}, {"author": "Tim Chown", "text": "<p>just wondering if discussions have happened about integrating. most people i know are too wary of using v1</p>", "time": "2025-07-21T13:22:53Z"}, {"author": "Matthieu Baerts", "text": "<p>Hi Neal, out of curiosity: are there any blocking points to upstream BBRv3 to the Linux kernel  -- other than the fact the BBRv3 draft is still being discussed?</p>", "time": "2025-07-21T13:27:08Z"}, {"author": "Christian Huitema", "text": "<p>Jitter sensitivity is a big issue with bandwidth measurements. Basically:<br>\nBW = bytes-ack / (delay +- jitter)</p>", "time": "2025-07-21T13:29:45Z"}, {"author": "Christian Huitema", "text": "<p>If the jitter is the same order magnitude as the RTT, this can cause huge deviations.</p>", "time": "2025-07-21T13:30:07Z"}, {"author": "Tim Chown", "text": "<p>Roland is at a CERN T1 site, could look at using perfSONAR for BBRv3 tests to other CERN experiment sites</p>", "time": "2025-07-21T13:42:55Z"}, {"author": "Lucas Pardue", "text": "<p>congestion_state_updated event <a href=\"https://datatracker.ietf.org/doc/html/draft-ietf-quic-qlog-quic-events.html#section-7.3\">https://datatracker.ietf.org/doc/html/draft-ietf-quic-qlog-quic-events.html#section-7.3</a></p>", "time": "2025-07-21T13:54:25Z"}]