From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-wm1-x341.google.com (mail-wm1-x341.google.com [IPv6:2a00:1450:4864:20::341]) (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 1C6973B29E for ; Tue, 9 Jul 2019 19:08:50 -0400 (EDT) Received: by mail-wm1-x341.google.com with SMTP id x15so404526wmj.3 for ; Tue, 09 Jul 2019 16:08:50 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20161025; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc:content-transfer-encoding; bh=RFGWIyJ8UomFdQHf8tvoMCGq2s8XebABzhj93IftLec=; b=oSIPvGeR+jIQOcc+MP/gZEPy/IxJhZ+c3VyVaFmD5QvwE12BEcR7NfB3SEYR/JYeKC HPXlwyq41dyQoPReVZs7+mBa1zpVbbImbzDAcRJGAmUmNys068D4iEaNbKknLG5R5Eoe JXJbbFF/PaZCwTIxqdN47cN0t57oWFe2RvHat6EYXSn/49dLHYtNuaL4otDlECWe8l1s ok5uxyNC97FzQAs0UqA0Fm3VrhYJ8NYiyNo2F5Movo3VyYcCgq6nWcp3GI9oBhaekVgE JbDE/nf1gTzdXGNiZDfyybnMz9DvLCnOJMPtnm7JV+FpDTuiPMrYqJ46wolDB+6L4C81 q83Q== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc:content-transfer-encoding; bh=RFGWIyJ8UomFdQHf8tvoMCGq2s8XebABzhj93IftLec=; b=Nwgkl4s4OOGIvP/1DNQ1P2S7TBA7tskQQWToBW2654QKqMMfhxZJQpKjiTcnJYMLuX afkE6l6jLzG+0uKKHgoaDPSuHHyp4QJp6X4+fg43/lwox71cKkEn5SUTVbuwXNs7pohG s3pk7UV6Fu4DfUGkqQrq70GErIVAcqRgCHf9oM+xi944beUWokJrqbVzxNH0BiJlhFBy il4nzLoNOKNlqPoRSA7umPS85vt9bJLYgK6S4BJEsMLeGLVqdMuPa5UpPTFU9heCwkI/ /QsqMowqxE0xa+kZR3XJ55gMVi9ipV1ZPq9zwQiCHiLl/ivMzbfqEBKGM7TMiqIbJJFr I3Uw== X-Gm-Message-State: APjAAAWVwhRI29rUY0hDH20mPmKXm3yDUNSIZb0U26xjs08DyL4Ygybe xL6tZlU6sGN6HSKuSZLAM589P2bmPv96YBMuieVv6g== X-Google-Smtp-Source: APXvYqxAwY5qqyM/Rm+/hPBY+3DaOVS6fwNKmfRjAE31RbwuxWAjMxzuFiEy223ru7PZRpb3Pl4HJWTFMyzrjBt3Zyw= X-Received: by 2002:a1c:c1c1:: with SMTP id r184mr1663655wmf.9.1562713728808; Tue, 09 Jul 2019 16:08:48 -0700 (PDT) MIME-Version: 1.0 References: <24f7b15a-129f-ca44-60e0-32c7d23eadf4@bobbriscoe.net> In-Reply-To: From: Yuchung Cheng Date: Tue, 9 Jul 2019 16:08:04 -0700 Message-ID: To: Jonathan Morton Cc: Bob Briscoe , tcpm IETF list , "ecn-sane@lists.bufferbloat.net" , tsvwg IETF list Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Mailman-Approved-At: Wed, 10 Jul 2019 05:05:17 -0400 Subject: Re: [Ecn-sane] [tsvwg] ECN CE that was ECT(0) incorrectly classified as L4S X-BeenThere: ecn-sane@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: Discussion of explicit congestion notification's impact on the Internet List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 09 Jul 2019 23:08:50 -0000 On Tue, Jul 9, 2019 at 8:41 AM Jonathan Morton wrot= e: > > > On 13 Jun, 2019, at 7:48 pm, Bob Briscoe wrote: > > > > 1. It is quite unusual to experience queuing at more than one > > bottleneck on the same path (the available capacities have to > > be identical). > > Following up on David Black's comments, I'd just like to note that the ab= ove is not the true criterion for multiple sequential queuing. > > Many existing TCP senders are unpaced (aside from ack-clocking), includin= g FreeBSD, resulting in potentially large line-rate bursts at the origin - = especially during slow-start. Even in congestion avoidance, each ack will = trigger a closely spaced packet pair (or sometimes a triplet). It is then = easy to imagine, or to build a testbed containing, an arbitrarily long sequ= ence of consecutively narrower links; upon entering each, the burst of pack= ets will briefly collect in a queue and then be paced out at the new rate. > > TCP pacing does largely eliminate these bursts when implemented correctly= . However, Linux' pacing and IW is specifically (and apparently deliberate= ly) set up to issue a 10-packet line-rate burst on startup. This effect ha= s shown up in SCE tests to the point where we had to patch this behaviour o= ut of the sending kernel to prevent an instant exit from slow-start. We (Google TCP folks) are internally experimenting (always) pacing IW. May hurt very long RTT and short transfers (<=3DIW), but could be an overall win. > > - Jonathan Morton >