Re: [tsvwg] ECT(1) Flag Day Plausibility

["Bless, Roland (TM)" <[e-mail]>]

Hi Jake,

On 15.05.21 at 00:05 Holland, Jake wrote:
> On 05-14, 5:35 AM, "Bless, Roland (TM)" <[e-mail]> wrote:
>> I find it somewhat odd to deprecate/revise a Proposed Standard in order
>> to rollout an Experiment.
> RFC 8311 also did this, I don't see this idea as especially different
> on this point.  I mean sure, maybe it's odd, but sometimes the right
> thing to do can be odd.
I agree.
>  From the poll at the interim, there seems to be significant support in
> the WG for moving this ahead regardless of the conflicts with existing
> BCPs, which is about what I expected based on past discussions.  The
> flag day proposal is just trying to look into one of the possibilities
> for making that align better.
Unfortunately, I missed the interim meeting.
>> Moreover, where can I find statistics about ECN usage along Internet
>> paths? I'm aware of
>> https://www.ietf.org/archive/id/draft-heist-tsvwg-ecn-deployment-observations-02.html,
>> but are there others?
> I did a presentation last year sharing a few observations I made:
> https://www.youtube.com/watch?v=1326B7YYwLM&t=4338s
> https://datatracker.ietf.org/meeting/interim-2020-maprg-01/materials/slides-interim-2020-maprg-01-sessa-latency-aqm-observations-on-the-internet-01.pdf
>
I missed that.
> Those slides reference another presentation with some other numbers
> collected differently:
> https://www.youtube.com/watch?v=wKDgVSMUvis&t=32m23s
> https://www.ietf.org/proceedings/98/slides/slides-98-maprg-tcp-ecn-experience-with-enabling-ecn-on-the-internet-padma-bhooma-00#page=12
>
> Outside Pete's writeup that's all I've seen published yet.

Thanks for the pointers!

>> I'd expect many home routers to be the bottleneck, so their potential
>> CE marks wouldn't be visible in the Internet, but merely the
>> corresponding ECE flags in TCP packets.
> That's what I did, yes.
>
> Apple's measurements were from agents on the receive side IIUC and
> I think they said included upload traffic, which I took to mean they
> took both CE and ECE that didn't trigger their heuristics as evidence
> of a CE-marking path.
>
> HTH.

Thanks a lot!

Regards,

  Roland