Re: [tsvwg] David Black (individual) on safety of L4S for the Internet

[Neal Cardwell <[e-mail]>]

On Fri, May 8, 2020 at 4:01 PM Black, David <[e-mail]> wrote:
> Current practice is not to mix DCTCP-like traffic (1/p-class congestion
> control) with TCP-like traffic (1/sqrt(p)-class congestion control, e.g.,
> NewReno) in the same queue because the DCTCP-like traffic outcompetes the
> TCP-like traffic to the point of starvation of the latter.

For my education, does someone have a pointer to the experiments
underlying that description of DCTCP outcompeting Reno/CUBIC? I'm not
aware of that being a fundamental property of DCTCP. The DCTCP
algorithm specifies a Reno-style response to loss -
https://tools.ietf.org/html/rfc8257#section-3.5 - and has Reno-style
increase functions as well. So the algorithm should be Reno friendly,
AFAIK.

The description of DCTCP outcompeting Reno reminds me of the behavior
of the Linux DCTCP implementation, before this 2019 bug fix by Koen De
Schepper and Olivier Tilmans:

  https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=aecfde23108b8e637d9f5c5e523b24fb97035dc3
  tcp: Ensure DCTCP reacts to losses

AFAIK, Linux DCTCP with this fix no longer outcompetes Reno.

> In contrast to reliance on TCP Prague, that combination of FQ and bleaching
> looks like it could provide a robust safety case.

AFAICT bleaching ECT(1) to Not-ECT should be sufficient for safety,
without also requiring FQ. An L4S flow that has  ECT(1) bleached to
Not-ECT should then be treated Like Reno by the bottleneck in
question. And, like DCTCP, the L4S specs specify a Reno-compatible
loss response for TCP Prague:
  https://tools.ietf.org/html/draft-ietf-tsvwg-ecn-l4s-id-09#appendix-A.1.3

So at least to a first approximation it seems like if an L4S flow is
bleached by an ISP to Not-ECT, is then treated by the ISP's
bottlenecks like non-ECN Reno, and then (per the L4S spec) behaves
like non-ECN Reno, L4S should be safe.

Apologies for the noise if I have missed something.

best regards,
neal