From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from z.eggo.org (z.eggo.org [80.235.105.138]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 9B88B3B260 for ; Mon, 6 Jun 2016 11:29:42 -0400 (EDT) Received: from localhost (localhost [127.0.0.1]) by z.eggo.org (Postfix) with ESMTP id 02C635800DD for ; Mon, 6 Jun 2016 18:29:41 +0300 (EEST) Received: from z.eggo.org ([127.0.0.1]) by localhost (z.eggo.org [127.0.0.1]) (amavisd-new, port 10032) with ESMTP id jGs6Xq7g3Ctp for ; Mon, 6 Jun 2016 18:29:40 +0300 (EEST) Received: from localhost (localhost [127.0.0.1]) by z.eggo.org (Postfix) with ESMTP id D7A10580117 for ; Mon, 6 Jun 2016 18:29:40 +0300 (EEST) X-Virus-Scanned: amavisd-new at harvee.org Received: from z.eggo.org ([127.0.0.1]) by localhost (z.eggo.org [127.0.0.1]) (amavisd-new, port 10026) with ESMTP id 9qG2bo5E03Fr for ; Mon, 6 Jun 2016 18:29:40 +0300 (EEST) Received: from [10.42.66.119] (173-14-129-9-NewEngland.hfc.comcastbusiness.net [173.14.129.9]) by z.eggo.org (Postfix) with ESMTPSA id 473C75800DD for ; Mon, 6 Jun 2016 18:29:40 +0300 (EEST) To: "cerowrt-devel@lists.bufferbloat.net" From: Eric Johansson Message-ID: <55fdf513-9c54-bea9-1f53-fe2c5229d7ba@eggo.org> Date: Mon, 6 Jun 2016 11:29:38 -0400 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:45.0) Gecko/20100101 Thunderbird/45.1.1 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="------------22E3A9075FD4BD088C9E9DC7" Subject: [Cerowrt-devel] trying to make sense of what switch vendors say wrt buffer bloat X-BeenThere: cerowrt-devel@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: Development issues regarding the cerowrt test router project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 06 Jun 2016 15:29:42 -0000 This is a multi-part message in MIME format. --------------22E3A9075FD4BD088C9E9DC7 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable one of my clients asked arista about bufferbloat issues in their switches. here was their response. is their analysis right? ------ Buffer bloat was a relevant on 10/100M switches, not 10Gb switches. At 10Gb we can empty the queue in ~100ms, which is less than the TCP retransmission timers, therefore no bloat. Buffer bloat can happen at slower speeds, but not an issue at the speeds we have on our switches.=20 There are some articles regarding bufferbloat on the net, but buffer bloat is not a problem on our switches. Some of the information regarding bufferbloat sites Internet routers where packets can be held in queues of very large buffers for several seconds, up to 10 seconds which can cause TCP retransmission problems and lower overall application performance when going across the public Internet. The buffers on the Arista 7500 are 128MB of packet buffers per 10GbE port coupled to a fully arbitrated (VOQ) virtual output queue forwarding system. At 10Gbps this is ~100msec of buffer capacity which is an order of magnitude from =E2=80=981 second=E2=80=99 and 2 orders of magnitude fr= om the 10 seconds worst case identified in buffer bloat documents. We (Arista) have switching systems with large buffers and high port count, or low buffers and high port count running one operating system. Buffer bloat is real in systems that would have more than 1.25GB of packet buffer per 10Gb port - none of these systems contribute to the buffer bloat issue. We position deep buffering switches where lossless performance is necessary. =20 --------------22E3A9075FD4BD088C9E9DC7 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable

one of my clients asked arista about bufferbloat issues in their switches.=C2=A0 here was their response= .=C2=A0 is their analysis right?

------

Buffer bloat was a relevant on 10/100M switches, not 10Gb switches. At 10Gb we can empty the queue in ~100ms, which is less than the TCP retransmission=C2=A0timers, therefore=C2=A0no bloat. Bu= ffer bloat can happen at slower speeds, but not an issue at the speeds we have on our switches.=C2=A0

There are some articles regarding bufferbloat on the net, but buffer bloat is not a problem on our switches.=C2=A0 Some of the information regarding bufferblo= at sites Internet routers where packets can be held in queues of very large buffers for several seconds, up to 10 seconds which can cause TCP retransmission problems and lower overall application performance when going across the public Internet.

The buffers on the Arista 7500 are 128MB of packet buffers per 10GbE port coupled to a fully arbitrated (VOQ) virtual output queue forwarding system.=C2=A0 At 10Gbps this is ~100msec of buffer capa= city which is an order of magnitude from =E2=80=981 second=E2= =80=99 and 2 orders of magnitude from the 10 seconds worst case identified in buffer bloat documents.

We (Arista) have switching systems with large buffers and high port count, or low buffers and high port count=C2=A0running one operating system. Buff= er bloat is real in systems that would have more than 1.25GB of packet buffer per 10Gb port - none of these systems contribute to the buffer bloat issue.

We position deep buffering switches where=C2=A0lossless performance is necessary.=C2=A0=C2= =A0

--------------22E3A9075FD4BD088C9E9DC7--