Re: [Rpm] [Bloat] Relentless congestion control for testing purposes

[Matt Mathis <mattmathis@google.com>]

[-- Attachment #1: Type: text/plain, Size: 1803 bytes --]

Please don't send this upstream.   It would makesTCP into the
evil transport from hell.    Modern loss recovery is robust enough to run
hardcoded cwnd=<blat> and persistent losses, Please don't make this too
easy for people who are intent on getting their "fair" share of the network
before the greedy people.

Dave overlooked an important detail in relentless TCP: It reduced the
window by exactly the losses, such that the presented load was exactly the
quantity of data successfully delivered on the previous RTT.   I have
forgotten the details of the increase function, but it was something Reno
like but only on loss less RTTs.



If you want to adapt TCP
Thanks,
--MM--
The best way to predict the future is to create it.  - Alan Kay

We must not tolerate intolerance;
       however our response must be carefully measured:
            too strong would be hypocritical and risks spiraling out of
control;
            too weak risks being mistaken for tacit approval.


On Fri, Oct 1, 2021 at 11:02 AM Luigi Rizzo via Rpm <
rpm@lists.bufferbloat.net> wrote:

> On Fri, Oct 1, 2021 at 6:33 PM Bob McMahon <bob.mcmahon@broadcom.com>
> wrote:
> >
> > hmm, this looks interesting to a test & measurement guy. Can it be done
> with a setsockopt? I might want to add this as an iperf2 option,
> particularly if it's broadly available,
>
>
> I would be happy to submit it as one or two upstream patches --
> perhaps one to implement
> the basic "ignore_holes" + setsockopt(), and another mechanism (if
> there isn't one
> already) to override defaults sockopts on certain sockets.
>
> I do think we need more readily available testing tool,
>
> cheers
> luigi
> _______________________________________________
> Rpm mailing list
> Rpm@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/rpm
>

[-- Attachment #2: Type: text/html, Size: 2756 bytes --]