Re: [nwcrg] [LOOPS] BOF co-chairs thinking on LOOPS next steps

Well if you to encode IP packets and you have encryption you will get into
the problem of the proposed coding scheme for QUIC that recovers  ALL
packets without knowing what the losses were due to and messing with
congestion control above or not knowing what these packets carry to
determine if they need coding or not or to define the adaptive schemes. Our
solution is before encryption because of that. And as Emmanuel said at
layer 4 you also need to deal with TCP congestion control unless you split
the sessions.

There is a lot of accumulated research, development and implementations out
there :) But ok you are talking to the right people!

mjm

Marie-José Montpetit, Ph.D.
Research Affiliate, MIT Media Laboratory
mariejose@mjmontpetit.com
mariejo@mit.edu

On July 30, 2019 at 1:46:51 PM, Spencer Dawkins at IETF (
spencerdawkins.ietf@gmail.com) wrote:

Hi, Emmanuel,

On Tue, Jul 30, 2019 at 10:14 AM Emmanuel Lochin <
emmanuel.lochin@isae-supaero.fr> wrote:

> Hi Spencer, all,
>
> On 30/07/2019 16:22, Spencer Dawkins at IETF wrote:
>
>
>    1. How a sender tunes FEC dynamically - is that automatic, based on
>    FEC mechanisms people are thinking about, or is there work to do there?
>    2. How a sender knows whether to do FEC, retransmission, or both FEC
>    and retransmission, dynamically?
>    3. How a sender knows that it shouldn't be doing anything, because
>    anything it does won't help ("first, do no harm")?
>
> IMHO, each answer depends on which layer you seek to apply FEC and why.
> Adding and adapting FEC at L2 or L3 does not have the same impact than at
> L4 and does not lead to the same adaptive mechanism. Then, adding FEC at L4
> also depends on whether you are congestion controlled or not, you want to
> stick to TCP-friendliness, your QoS objectives, and many other things...
> Finally concerning #3, I would say that monitoring the e2e delay would
> give you a good indication to enable FEC or not. If the delay is "small",
> you won't beat ARQ.
>

My understanding from discussions around the LOOPS BOF is that this is
intended to be transport-agnostic, so most likely L3. I haven't heard
anyone talking about L2 LOOPS, but as you note, that depends on the
networks that you're trying to help out.

Thanks for the insight! (I thought that was true, but I'm not an NWC expert
by any stretch)

Spencer

> Emmanuel
>
>
> --
> Emmanuel LOCHIN
> Professeur ISAE
> ISAE SUPAERO - Institut Supérieur de l'Aéronautique et de l'Espace
> 10 avenue Edouard Belin - BP 54032 - 31055 TOULOUSE CEDEX 4 FRANCE -http://www.isae-supaero.fr
> Tel +33 5 61 33 84 85 - Fax (+33) 5 61 33 83 45
> Web : http://personnel.isae.fr/emmanuel-lochin/
>
> --
> LOOPS mailing list
> LOOPS@ietf.org
> https://www.ietf.org/mailman/listinfo/loops
>
-- 
LOOPS mailing list
LOOPS@ietf.org
https://www.ietf.org/mailman/listinfo/loops