Hi Ted,
This document aims at offering guidance on the use of TCP in IoT
environments or use cases, based on contributions from a really long list
of folks from the LWIG and TCPM working groups, with diverse backgrounds
and expertise (from academia, industry, implementers, etc.). If
interested, you may want to see the Acknowledgments section.
Could you please provide any pointers to "existing research that
shows that MSS of greater than perhaps five lowpan frames is quite
harmful." ?
Also, we believe that the statement that the document "recommends larger
mss" is not true or accurate. In 4.1.1 we rather discuss the benefits of
not exceeding an MTU of 1280 bytes, and we give guidelines that may help
in this regard.
There is also the last paragraph of 4.1.1, which includes:
"using larger MSS (to a suitable extent) may be beneficial, especially
when transferring large payloads, as it reduces the number of packets
(and packet headers) required for a given payload"
but it purposefully includes "to a suitable extent", which is also not
quantified, and it is given in the context of transferring large payloads.
And the "may" is also very relevant here.
We believe that what that last paragraph states is actually quite obvious,
implicitly assuming that the application of that paragraph will be done in
a careful, cautious way...
Best regards,
Carles (on behalf of all coauthors)
> On Sep 16, 2020, at 10:01 AM, Ted Lemon <mellon@xxxxxxxxx> wrote:
>> To a first approximation, this document is harmful garbage. It
>> recommends larger mss. Probably not based on opex.
>
> Wow, my first reply-all screwup in ages, and it was pretty bad. Really
> sorry about thisâ??I did not intend to be this transparent on a public
> mailing list, and I hope my first impression is wrong. But with that said,
> that is my first impression, and I will be giving the document further
> review. Itâ??s very disappointing to me that lwig has come out with this
> without, as far as I can tell, any reference to existing research that
> shows that MSS of greater than perhaps five lowpan frames is quite
> harmful.
>
> Could the authors common on what operational experience this document is
> based on?
>
> _______________________________________________
> Lwip mailing list
> Lwip@xxxxxxxx
> https://www.ietf.org/mailman/listinfo/lwip
>
--
last-call mailing list
last-call@xxxxxxxx
https://www.ietf.org/mailman/listinfo/last-call
