From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mout.gmx.net (mout.gmx.net [212.227.17.21]) (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 85A5C3B29D for ; Sun, 3 May 2020 10:43:01 -0400 (EDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=gmx.net; s=badeba3b8450; t=1588516980; bh=cMifTzznJTqU6+AoQRin5tYq/9Gj8q93AM1d97OC6v0=; h=X-UI-Sender-Class:Date:In-Reply-To:References:Subject:To:CC:From; b=ch/QfxrOppTUgD3012ftNjq9N44i31CkIqdKRItlCF8xOzpiu2JNQndosxvx8ivGS ImN5lsJVq6MB5tsumvWGR9Mz1fGfaqfB5z4Gg9Kvz9XiFV9JVPhgHvMbYjc5YyGQ9b L2foZwb9vLWdqvbjpq51OIf6tyCP9PkGIhDk9Vp0= X-UI-Sender-Class: 01bb95c1-4bf8-414a-932a-4f6e2808ef9c Received: from [192.168.42.160] ([95.112.125.18]) by mail.gmx.com (mrgmx104 [212.227.17.168]) with ESMTPSA (Nemesis) id 1MmULx-1inQkk3qzZ-00iTvd; Sun, 03 May 2020 16:43:00 +0200 Date: Sun, 03 May 2020 16:42:58 +0200 User-Agent: K-9 Mail for Android In-Reply-To: References: <3764C024-5F18-41C8-98F0-9FA423ED5860@gmail.com> MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----0C4XB6Y0ASKKNH29MXHOFBFOBQVDYI" Content-Transfer-Encoding: 7bit To: bloat@lists.bufferbloat.net, Dave Taht , Daniel Sterling CC: Rich Brown ,bloat From: Sebastian Moeller Message-ID: X-Provags-ID: V03:K1:JtzxROdx2JVSxOV8J4kEqL1sDPJ6pXc7ssVlimRHjcVIbbwW1XM hF0C8mz+IlSsPAy4j2PzWnFxrNQJpv6LclfoRl2Qm1TAqT6O/I/XrM4ghqmWAPOyGF2s+Pg 2aRwQxlNjojzWq4uEV4Klwd24GX0gx1UCr97oZA5Meg8Qys0J48k7TH6HdoaF3Cye5G+J/b 0K2yXtzuGAGBGzpGUqOQg== X-Spam-Flag: NO X-UI-Out-Filterresults: notjunk:1;V03:K0:5SleEkHsvBg=:sovBFJiSjXKtNDF1A+cOfk x2FocyYByDSTIfU+vFOHu6GcDYVLKkCkxNmqDrc9pNCP6Xi7+eVuPxcDongxz+IQMfiS4Q0LA jykdO4uxyo8+cQsaWuyOc15rkL9M08tf+vNY8jDUdOmRjC2F4UC62JdadC17JC+pW2ndgv9R9 OIT3DP1Edu8SMYBhp4tmOHwtb/cZwe96k8k2aMg9w+x5IqRuD0gTf0E2NxIFx6+dHpOPzdJFi k4bsmnM1Fs7ePaEJmb48HO0OuFJngDL1fMKfJFOfkTwdvcR80djqJsLiKNa9Zx1dU09ce+M5b pu4h28w6pEwd75jGM1F15XKoaDSfRY5S+vnU8SmssseqWuWjaPRn70KDqImCGY0Yj0VFAbYK8 yhijWqS95m8TeTxNh6+8AJENxYWV38ih1QtS1ic5ghyVk4KzELQR54eqXBNhWUbKUZaQ+En1R M5p1oTL7eH87D+lo4G3szZOXTRYJexqWWEdTt7hLulT8iH9gAdf9op1MXxgxvK61X1Mdk0j8W 8twyo2rAwIkFHdqv3WpBrpFN0tzpV3Dac0/zMB3ApzghoLu3oQ/SZdW8fgPwKk/ljEO8T/r7Q WvtfjCAXR4zkpUCxEzHPC8Dbtwxuxy3jy0g06kpZvBycpNtv3gFPItKQrdheRMoMNAumoW5LJ iuwd8agnxOT2ldQqR0Zyu364CgK+VFAT2ZDmTIrjx7HAz6xMfbWjD0fbG8yZDTi/RES8UuFxE 5Y6a2+5H0G/V8Q8Jy2U93Xi1Mm/WZl6TvUz17moSNXWKD+KkFLiCtd4DwhCI2LuIlWW7rOS3T szwYSYMo1EoAYW9X+eg+6IfZM1p1U2UdqpQzemUX86wMENmrSeO5VRMISeWrT2FjKWITmTrO2 mbptx+Fx69Fxnme162mUTZ7A+JQarF9OlMUqYEon3xunj3b9d4Tea5OAXlZvszY5BqMklHG0I Spy28OnnZCGVr3DJeV4mrAN7cy5P9VJfCvxxOMnBNtoZOmeLmdUNnCKpk1Cfi0DuYVebdch7x oo47E9vQt3MNk4Z1DgZ3QiElu1j2o90dBqR+7NWRszUSCvtFgE79dhoxe/85L6buVS2nXyWDR hlY13OwyUYOl0kvrTyLKOs0ND2S/0yMEMzne1YgV3fFSTA9FYo3Hs5Fowxj32XeKWURx75BHm rMvf3aHHiuAHz0yH89drlmzzP7tr5sqr5AtINj2ZjpyoqZGvlfHAivSIA2JDf5AC6JaFz35O5 2yKw6QHog3vI+CBmY Subject: Re: [Bloat] Multiple WAN ports & SQM? X-BeenThere: bloat@lists.bufferbloat.net X-Mailman-Version: 2.1.20 Precedence: list List-Id: General list for discussing Bufferbloat List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 03 May 2020 14:43:01 -0000 ------0C4XB6Y0ASKKNH29MXHOFBFOBQVDYI Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable I guess the question is, what Rich needs more urgently, more aggregate rate= or more single-flow performance? Then for bonding one needs a dedicated head-end device on the internet sid= e of things, while mwan3 on the router alone should work with any independe= nt links for failover and load sharing, IIRC=2E Best Regards Sebastian On 3 May 2020 16:33:56 CEST, Dave Taht wrote: >not huge on bonding, simpler to just get the two uplinks and split >flows across them with an sqm instance for each and a tc hash >directing flows at one or another=2E > >On Sun, May 3, 2020 at 7:30 AM Daniel Sterling > wrote: >> >> When I had both DSL and cable modem, I compiled Linux with this patch >set to make multi gateway NAT work and it worked great >> >> http://ja=2Essi=2Ebg/#routes >> >> Should be able to use that plus ifb+cake on each NIC to do the right >thing, aye? >> >> As an aside, I'm kind of furious that NAT fix never got merged >upstream :( it's so useful for multiple uplinks >> >> -- Dan >> >> On May 3, 2020, at 10:23 AM, Rich Brown >wrote: >> >> =EF=BB=BFGiven the crummy internet service in my area (DSL, max of >15mbps/1mbps), I wonder if we could improve things by getting a second >connection from our ISP and "bonding" the two links together in my >OpenWrt router=2E >> >> I see both Multiwan (which is self-described as old) and mwan3=2E >> >> But neither would seem to offer the kinds of latency control >(SQM/fq_codel/cake) that the cool kids in networking have come to >expect=2E >> >> Any recommendations from this group for such an effort? Thanks=2E >> >> Rich >> _______________________________________________ >> Bloat mailing list >> Bloat@lists=2Ebufferbloat=2Enet >> https://lists=2Ebufferbloat=2Enet/listinfo/bloat >> >> _______________________________________________ >> Bloat mailing list >> Bloat@lists=2Ebufferbloat=2Enet >> https://lists=2Ebufferbloat=2Enet/listinfo/bloat > > > >--=20 >Make Music, Not War > >Dave T=C3=A4ht >CTO, TekLibre, LLC >http://www=2Eteklibre=2Ecom >Tel: 1-831-435-0729 >_______________________________________________ >Bloat mailing list >Bloat@lists=2Ebufferbloat=2Enet >https://lists=2Ebufferbloat=2Enet/listinfo/bloat --=20 Sent from my Android device with K-9 Mail=2E Please excuse my brevity=2E ------0C4XB6Y0ASKKNH29MXHOFBFOBQVDYI Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable I guess the question is, what Rich needs more urge= ntly, more aggregate rate or more single-flow performance?
Then for bond= ing one needs a dedicated head-end device on the internet side of things, w= hile mwan3 on the router alone should work with any independent links for f= ailover and load sharing, IIRC=2E

Best Regards
Sebastian<= br>
On 3 May 2020 16:33:56 CEST, Dave Taht &l= t;dave=2Etaht@gmail=2Ecom> wrote:
not huge on bonding, simpler to just get the two upl=
inks and split
flows across them with an sqm instance for each and a tc =
hash
directing flows at one or another=2E

On Sun, May 3, 2020 at =
7:30 AM Daniel Sterling
<sterling=2Edaniel@gmail=2Ecom> wrote:
=
>
 When I had both D=
SL and cable modem, I compiled Linux with this patch set to make multi gate=
way NAT work and it worked great

 http://ja=2Essi=2Ebg/#routes

 Should be able to use that =
plus ifb+cake on each NIC to do the right thing, aye?

 As an aside, =
I'm kind of furious that NAT fix never got merged upstream :( it's so usefu=
l for multiple uplinks

 -- Dan

 On May 3, 2020, at 10:23 AM, =
Rich Brown <richb=2Ehanover@gmail=2Ecom> wrote:

 =EF=BB=BFGive=
n the crummy internet service in my area (DSL, max of 15mbps/1mbps), I wond=
er if we could improve things by getting a second connection from our ISP a=
nd "bonding" the two links together in my OpenWrt router=2E

 I see b=
oth Multiwan (which is self-described as old) and mwan3=2E

 But neit=
her would seem to offer the kinds of latency control (SQM/fq_codel/cake) th=
at the cool kids in networking have come to expect=2E

 Any recommend=
ations from this group for such an effort? Thanks=2E

 Rich
 Bloat=
 mailing list
 Bloat@lists=2Ebufferbloat=2Enet
 https://lists=2Ebufferbloat=2Enet/l=
istinfo/bloat
 Bloat mailing list
 Bloat@lists=2Ebufferbloat=2Ene=
t
 https://=
lists=2Ebufferbloat=2Enet/listinfo/bloat


=

--
Sent from my Android device with K-9 Mail=2E = Please excuse my brevity=2E ------0C4XB6Y0ASKKNH29MXHOFBFOBQVDYI--