From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from smtp122.iad3a.emailsrvr.com (smtp122.iad3a.emailsrvr.com [173.203.187.122]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by lists.bufferbloat.net (Postfix) with ESMTPS id 345AC3CB37 for ; Sun, 7 May 2023 18:22:34 -0400 (EDT) Received: from app55.wa-webapps.iad3a (relay-webapps.rsapps.net [172.27.255.140]) by smtp32.relay.iad3a.emailsrvr.com (SMTP Server) with ESMTP id B44511988; Sun, 7 May 2023 18:22:33 -0400 (EDT) Received: from deepplum.com (localhost.localdomain [127.0.0.1]) by app55.wa-webapps.iad3a (Postfix) with ESMTP id 99D4F62294; Sun, 7 May 2023 18:22:33 -0400 (EDT) Received: by apps.rackspace.com (Authenticated sender: dpreed@deepplum.com, from: dpreed@deepplum.com) with HTTP; Sun, 7 May 2023 18:22:33 -0400 (EDT) X-Auth-ID: dpreed@deepplum.com Date: Sun, 7 May 2023 18:22:33 -0400 (EDT) From: "David P. Reed" To: "Michael Richardson" Cc: "Dave Taht" , "cerowrt-devel" MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_20230507182233000000_79343" Importance: Normal X-Priority: 3 (Normal) X-Type: html In-Reply-To: <29691.1683477385@localhost> References: <19253.1683475762@localhost> <29691.1683477385@localhost> X-Client-IP: 209.6.168.128 Message-ID: <1683498153.625631572@apps.rackspace.com> X-Mailer: webmail/19.0.23-RC X-Classification-ID: cedda5f6-a0a3-475d-b3d1-21588891ba46-1-1 Subject: Re: [Cerowrt-devel] risc-v 2 ethernet port router 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: Sun, 07 May 2023 22:22:34 -0000 ------=_20230507182233000000_79343 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable =0ACelerons and Atoms have been doing just fine with fq_codel on 1 GigE. Ho= wever, the prices for 2 port boxes seems to be about $200 with RAM and enou= gh flash to boot LInux.=0A =0AI personally think the time has come to start= using 2.5 GigE on the typical 2 port home routers, and avoid USB, as it ju= st makes support harder - use PCIe.=0A =0ABut I'm a special case, my home l= ab and all of my home is 10 GigE backbone, with 2.5 GigE to the wireless AP= s. This calendar year, I'm moving to either 25 or 40 GigE on the backbone a= nd in the lab, but I haven't picked the final setup. Partly it depends on w= hat I can get from my external ISP for a reasonable price - hopefully 2 Gig= or better will come here, we have many competitors (3 cable providers and = various small business fixed wireless).=0A =0AI'm wondering about how fq_co= del will handle these higher speeds on inexpensive computers.=0A =0ADavid= =0A =0AOn Sunday, May 7, 2023 12:36pm, "Michael Richardson via Cerowrt-deve= l" said:=0A=0A=0A=0A> _______________= ________________________________=0A> Cerowrt-devel mailing list=0A> Cerowrt= -devel@lists.bufferbloat.net=0A> https://lists.bufferbloat.net/listinfo/cer= owrt-devel=0A> =0A> Dave Taht wrote:=0A> > Well, I do= not know if it is fq_codel capable. A lot of folk have been=0A> > missing = on adding in BQL support of late, and as to whether it has the=0A> =0A> I m= ean... at $119, if it can do fq_codel, then the price is totally=0A> accept= able :-)=0A> =0A> > horsepower to do SQM at these rates is undetermined - I= have generally=0A> > felt gobs of cache were required to do soft rate shap= ing.=0A> =0A> > Secondly, you can always add a 3rd ethernet port via usb no= wadays.=0A> =0A> Yes, it's just less elegant if you want to have 5-6 of the= m in a lab.=0A> =0A> ------=_20230507182233000000_79343 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable

Celerons and Atoms hav= e been doing just fine with fq_codel on 1 GigE. However, the prices for 2 p= ort boxes seems to be about $200 with RAM and enough flash to boot LInux.=0A

 

=0A

I personally th= ink the time has come to start using 2.5 GigE on the typical 2 port home ro= uters, and avoid USB, as it just makes support harder - use PCIe.

=0A

 

=0A

But I'm a special case,= my home lab and all of my home is 10 GigE backbone, with 2.5 GigE to the w= ireless APs. This calendar year, I'm moving to either 25 or 40 GigE on the = backbone and in the lab, but I haven't picked the final setup. Partly it de= pends on what I can get from my external ISP for a reasonable price - hopef= ully 2 Gig or better will come here, we have many competitors (3 cable prov= iders and various small business fixed wireless).

=0A

I'm wondering about how fq_codel will h= andle these higher speeds on inexpensive computers.

=0A

 

=0A

David

=0A

&nbs= p;

=0A

On Sunday, May 7, 2023 12:36pm, "Michael Rich= ardson via Cerowrt-devel" <cerowrt-devel@lists.bufferbloat.net> said:=

=0A
=0A

> _______________________________________________
> Cerowrt-dev= el mailing list
> Cerowrt-devel@lists.bufferbloat.net
> htt= ps://lists.bufferbloat.net/listinfo/cerowrt-devel
>
> Dave= Taht <dave.taht@gmail.com> wrote:
> > Well, I do not know= if it is fq_codel capable. A lot of folk have been
> > missing = on adding in BQL support of late, and as to whether it has the
> > I mean... at $119, if it can do fq_codel, then the price is totall= y
> acceptable :-)
>
> > horsepower to do SQM a= t these rates is undetermined - I have generally
> > felt gobs o= f cache were required to do soft rate shaping.
>
> > Se= condly, you can always add a 3rd ethernet port via usb nowadays.
> =
> Yes, it's just less elegant if you want to have 5-6 of them in a= lab.
>
>

=0A
 ------=_20230507182233000000_79343--