Re: [tsvwg] start of WGLC on L4S drafts

[Dave Taht <[e-mail]>]

As one of the contributors to codel, pie, fq-codel, cake, and fq-pie
in linux, openwrt, ns2, and ns3, and as the co-founder of the
Bufferbloat Project, I have re-read the L4S drafts, and do not support
forwarding them on to the IESG.

The experiment(s) are over. The technical evidence against the core
L4S approaches is overwhelming. Rather than proceed to any RFC status
these documents should depart into the same memory hole as source
quench did.

Rather than re-iterate technical point after technical point here...
if, in this last call, these drafts are approved to send to the IESG,
I request that the entirety of the SCE related data[1] be attached as
an appendix, much as Richard Feynman affixed his Appendix A to the
Challenger Commission report, lest this be the last triumph of process
and politics over sound science and engineering in the IETF.

In the long run, existing transport responses to RFC3168-style ECN
handling should be improved, and the more permissive aspects of
RFC8311 repudiated, in favor of a consistent, well understood, IP
layer and internet wide, sharp congestion response to a single CE
mark, as is the default, today.

[1] https://github.com/heistp

-- 
Fixing Starlink's Latencies: https://www.youtube.com/watch?v=c9gLo6Xrwgw
Dave Täht CEO, TekLibre, LLC