Re: [tcpm] [iccrg] Cwnd growth after fast convergence in CUBIC

Yes, the RFC is different from the Linux Cubic in case of fast convergence.

I would suggest changing Equation (2) of RFC 8312 to
K = cubic_root((W_max* - cwnd)/C) (Eq. 2)

where cwnd is the cwnd at the beginning of a congestion avoidance
stage. That is, W_cubic(0)=the cwnd at the beginning of a congestion
avoidance stage

By the way, sometime the fast recovery stage also changes cwnd based
on the number of packets in flight. As a result the cwnd at the
beginning of a congestion avoidance stage may be different from
W_last_max*beta_cubic.

Therefore, I would prefer to using the following Equation (2) in RFC 8312

K = cubic_root((W_max* - cwnd)/C) (Eq. 2)

instead of

K = cubic_root(W_max - W_last_max*beta_cubic)/C) (Eq. 2)

Thank you all

Lisong

On Fri, Nov 6, 2020 at 8:11 AM Neal Cardwell <ncardwell=
40google.com@dmarc.ietf.org> wrote:

> On Fri, Nov 6, 2020 at 3:28 AM Lars Eggert <lars@eggert.org> wrote:
> > That said, the intent was to document what the Linux code was doing
> > at the time. Is there a difference between the RFC and what Linux
> > has been doing or is doing now? If yes, it might be time to revise
> > the RFC.
>
> Yes, it seems that this is a case where the RFC differs from the Linux
> TCP CUBIC code.
>
> The RFC says:
>       K = cubic_root(W_max*(1-beta_cubic)/C) (Eq. 2)
>
> which is equivalent to:
>       K = cubic_root((W_max - W_max*beta_cubic)/C).  [B]
>
> The Linux code does something different: it computes K once per epoch
> immediately after the fast recovery or ECN cwnd reduction phase
> completes, using:
>
>       K = cubic_root((W_max - cwnd)/C).  [C]
>
> See the computation of bic_K here:
>   https://github.com/torvalds/linux/blob/master/net/ipv4/tcp_cubic.c#L254
>
> For the non-fast-convergence case, the result of [B] and [C] are
> equivalent, but for the fast convergence case they are not. So it
> looks like in the case of fast convergence the RFC formula produces a
> result that is incorrect, and does not match the Linux TCP CUBIC
> behavior.
>
> AFAICT,  changing Eq 2 to the following might fix the RFC and make it
> match the Linux TCP CUBIC behavior:
>
>       K = cubic_root(W_max - W_last_max*beta_cubic)/C) (Eq. 2)
>
> best,
> neal
>
> _______________________________________________
> iccrg mailing list
> iccrg@irtf.org
> https://www.irtf.org/mailman/listinfo/iccrg
>