Re: [L4s-discuss] BBRv3 and L4S

Neal Cardwell <ncardwell@google.com> Tue, 16 January 2024 16:47 UTC

MIME-Version: 1.0
References: <AM8PR07MB8137FCFD1AB8A5123D85062BC2732@AM8PR07MB8137.eurprd07.prod.outlook.com> <CADVnQykPM1iMmKmf-Pip6UORQ_vFQfNf25RQpxtNhW5pr4JHsA@mail.gmail.com>
In-Reply-To: <CADVnQykPM1iMmKmf-Pip6UORQ_vFQfNf25RQpxtNhW5pr4JHsA@mail.gmail.com>
From: Neal Cardwell <ncardwell@google.com>
Date: Tue, 16 Jan 2024 11:46:37 -0500
Message-ID: <CADVnQymnY+UZHK9Pfm8RwGo2ri9gM7hwXR0kXB7fCQiRVJB8_w@mail.gmail.com>
To: l4s-discuss@ietf.org, Ingemar Johansson S <ingemar.s.johansson@ericsson.com>
Content-Type: multipart/alternative; boundary="000000000000c95a8f060f12e033"
Archived-At: <https://mailarchive.ietf.org/arch/msg/l4s-discuss/H_sPMSEUDJwN0bKousTeYUmh6eU>
Subject: Re: [L4s-discuss] BBRv3 and L4S
Precedence: list

Hi Ingemar,

> Tried to figure out if BBRv3 supports L4S

In principle one could use BBRv3 in an L4S environment. However, its ECN
logic was designed for and is currently used only in
datacenter environments, and is not yet designed/tuned for L4S
environments. Specifically:

(1) The ECN logic in BBRv3 has only been tuned/used in production in
low-RTT datacenter environments, and is known to have shortcomings at
longer RTTs.

(2) The ECN logic in BBRv3 is not designed for good coexistence with Prague
CC, and would tend to starve Prague CC.

> and, if that is the case, does it use accurate ECN  ?

Linux TCP BBRv3 can use whatever ECN information the core TCP stack
provides: that can either be (1) DCTCP-style ECN signals echoed in ACKs
(that's how we use it internally within Google) or (2) Accurate ECN.

best regards,

neal

On Tue, Jan 16, 2024 at 9:52 AM Ingemar Johansson S <
ingemar.s.johansson@ericsson.com> wrote:

> Hi
>
>
>
> I am trying to catch up on BBRv3, found a presenation from IETF 117
>
>
> https://datatracker.ietf.org/meeting/117/materials/slides-117-ccwg-bbrv3-algorithm-bug-fixes-and-public-internet-deployment-00
>
> Tried to figure out if BBRv3 supports L4S and, if that is the case, does
> it use accurate ECN  ?
>
>
>
> Any information is greatly appreciated.
>
>
>
> /Ingemar
>
> =================================
>
> Ingemar Johansson  M.Sc.
>
> Master Researcher
>
>
>
> Ericsson AB
>
> GFTL ER Networks RNS Protocol & E2E Perf
>
> Labratoriegränd 11
>
> 977 53, Luleå, Sweden
>
> Phone +46-1071 43042
>
> SMS/MMS +46-73 078 3289 <+46%2073%20078%2032%2089>
>
> ingemar.s.johansson@ericsson.com
>
> www.ericsson.com
>
>
>
>          *The right to be ridiculous is *
>
>               *something I hold dear...*
>                               U2
>
> =================================
>
>
>

Re: [L4s-discuss] BBRv3 and L4S Neal Cardwell