From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-ed1-x52e.google.com (mail-ed1-x52e.google.com [IPv6:2a00:1450:4864:20::52e]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 790AE3B2A4 for ; Mon, 3 Jan 2022 19:24:04 -0500 (EST) Received: by mail-ed1-x52e.google.com with SMTP id z9so71831399edm.10 for ; Mon, 03 Jan 2022 16:24:04 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=broadcom.com; s=google; h=mime-version:from:date:message-id:subject:to:cc; bh=7MvVJW39dtaXvyHT/Tf/Or75A5XZAVpYBwZZsYi8e00=; b=X3jLY9PVBPU82Y+PvCRyufCTjGlaumv8zCuzC0C/xNk6NpXnmD5sF+2MQrnjT/F9PP 34ojd1RQ+CBzD6O63fDlAerRTfccvw5yf94GASU295Mpc6k5p3c5S072TSE6GKf8OSYi 3qGQGvKi2KCDV46RHCxffRnQh3cS+i9VDm26Q= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:from:date:message-id:subject:to:cc; bh=7MvVJW39dtaXvyHT/Tf/Or75A5XZAVpYBwZZsYi8e00=; b=ZY2BJPiH/4f3RA6osf+Dul2euiAvMypEZe5cIDEJa2MSI5n9W9pN3OcqGiFISsoeEQ oEqe59chKUg2iEX0cMzI4Dl8lY/YTIsQZdrTpCnIE15WH7srI3iDGDFaUPH8qPnehxSW gFIWj98uHSgtH+I5znQHjkHBgYfvcD+/9TggHgICY+XrNGpUOFIrAN02G1mlbH7uZYQD hPe8OmPneC0W/urwZLarytjjwXnc8UKh16wvC14AIztR+USndYD9QD+YmsRn5uFIVsnF FA7KmWGZFaU2YcMfgTJv9VUNJCndPnfhY7DdxaTb6X2CNwO9RVK4CHcry+eKqhKGkSmb djDQ== X-Gm-Message-State: AOAM530k6jNv+EP3coaqK7YV2UiGbSUYI9oFXNcRl0rBoNjB11WLK49d BftdJdTf92mUcXdUq2IsVJd11exPIyBHH2rNSZ1o/+pGmqPWWfbqAHGENVd2FWLaB0CEF7n4gIy XxCd0tYhkRBsEQFp+2rSKLaRg2AQnbbFsVnptezSGZGoOiW8= X-Google-Smtp-Source: ABdhPJzfC6f8ai5mltLhu9RPQfA5MLTdPzU8Laky7YNKpEQ1+ASpxbfvbhKIKo9vL5Su+//8mSVKs03OekgMTT3xqmM= X-Received: by 2002:a17:906:414c:: with SMTP id l12mr38404807ejk.641.1641255843145; Mon, 03 Jan 2022 16:24:03 -0800 (PST) MIME-Version: 1.0 From: Bob McMahon Date: Mon, 3 Jan 2022 16:23:52 -0800 Message-ID: To: Make-Wifi-fast Cc: Vint Cerf , Leonard Kleinrock , "David P. Reed" , dickroy@alum.mit.edu, Dave Taht , Neal Cardwell , Stuart Cheshire , Christoph Paasch , Matt Mathis Content-Type: multipart/alternative; boundary="00000000000084576405d4b6a780" Subject: [Make-wifi-fast] Network telemetry and non-parametric distributions X-BeenThere: make-wifi-fast@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 04 Jan 2022 00:24:04 -0000 --00000000000084576405d4b6a780 Content-Type: text/plain; charset="UTF-8" Hi All, Happy New Year! My apologies up front if this email is not of interest and hence spam to you. I was hoping to get some technical feedback from industry networking and queuing experts with respect to latency fault detection and forwarding path decisions. Iperf 2 flows code produces dendrograms (ward clustering) with respect to e2e latencies. (Latencies can be video frames, write to reads, packets, etc.) The metric used to populate the distance matrices comes per the Kolmogorov Smirnov two sample test which supports comparing non-parametric distributions. WiFi 6 plans to use MU transmissions to reduce latency as one TXOP or scheduled trigger can TX multiple AP/STA packets, i.e. kinda a form of multicast (just doesn't use L2 MCAST DAs but rather does it in the RF domain.) I'm wondering about realtime detection of "latency timing violations," and possibly using ML to identify better low latency path trees (LLPT) analogous to IP multicast RPT and SPT. The idea is to cluster the graphs in a way that packets get probabilistically bunched into microbursts per the final AP's MU groups (since it's assumed the last hop is WiFi.) Does this seem technically reasonable and, if so, is there a reasonable return on the engineering? I'd like to prototype it with iperf 2, python (pyflows) and on some test rigs but it's a lot of work and I don't want to take on that work if the return on engineering (ROE) is near zero. Thanks in advance for your thoughts and opinions, Bob -- This electronic communication and the information and any files transmitted with it, or attached to it, are confidential and are intended solely for the use of the individual or entity to whom it is addressed and may contain information that is confidential, legally privileged, protected by privacy laws, or otherwise restricted from disclosure to anyone else. If you are not the intended recipient or the person responsible for delivering the e-mail to the intended recipient, you are hereby notified that any use, copying, distributing, dissemination, forwarding, printing, or copying of this e-mail is strictly prohibited. If you received this e-mail in error, please return the e-mail to the sender, delete it from your computer, and destroy any printed copy of it. --00000000000084576405d4b6a780 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
Hi All,

Happy New Year! My apologies up front if th= is=C2=A0email=C2=A0is not of interest and=C2=A0hence spam to you.

I = was hoping to get some technical feedback from industry networking and queu= ing=C2=A0experts with respect to latency fault detection and forwarding pat= h decisions. Iperf 2 flows code produces dendrograms (ward clustering) with respect to e2e la= tencies. (Latencies can be video frames, write to reads, packets, etc.) The= metric used to populate the distance matrices comes per the Kolmogorov Smir= nov two sample test which supports comparing non-parametric distributio= ns.

WiFi 6 plans to use MU transmissions to reduce latency as one TX= OP or scheduled trigger can TX multiple AP/STA packets, i.e. kinda a form o= f multicast (just doesn't use L2 MCAST DAs but rather does it in the RF= domain.)

I'm wondering=C2=A0about realtime detection of "l= atency timing violations," and possibly=C2=A0using ML to identify bett= er low latency path trees (LLPT) analogous=C2=A0to IP multicast RPT and SPT= . The idea is to cluster the graphs in a way that packets get probabilistic= ally bunched into microbursts per the final AP's MU groups (since it= 9;s assumed the last hop is WiFi.)

Does this seem technically reason= able and, if so, is there a reasonable=C2=A0return on the engineering? I= 9;d like to prototype it with iperf 2, python (pyflows) and on some test ri= gs but it's a lot of work and I don't want to take on that work if = the return on engineering (ROE) is near zero.

Thanks in advance for = your thoughts and opinions,
Bob

This ele= ctronic communication and the information and any files transmitted with it= , or attached to it, are confidential and are intended solely for the use o= f the individual or entity to whom it is addressed and may contain informat= ion that is confidential, legally privileged, protected by privacy laws, or= otherwise restricted from disclosure to anyone else. If you are not the in= tended recipient or the person responsible for delivering the e-mail to the= intended recipient, you are hereby notified that any use, copying, distrib= uting, dissemination, forwarding, printing, or copying of this e-mail is st= rictly prohibited. If you received this e-mail in error, please return the = e-mail to the sender, delete it from your computer, and destroy any printed= copy of it. --00000000000084576405d4b6a780--