[Bloat] a foundation for flent?

Dave Taht dave.taht at gmail.com
Wed May 11 21:55:39 EDT 2022


All:

I started this proposal two weeks back in an attempt to understand and
apply for support from the NSF "POSE" grants program. It's due
tomorrow, but there's no way I'll make that deadline. There's a phase
II proposal that could be put together by October if the right people
can be found, and there are other options like an DoE SBIR or
commercial investment, etc that could also be pursued.

I felt, in drafting this, that with a major investment into flents PR,
and outreach, and ease of use, we'd be able to train more folk about
ways to fix more networks. If anyone would like to comment, it's here.
I liked to think the rhetoric and the accomplishments we had so far
with toke's wonderful tool, should speak for themselves.

https://docs.google.com/document/d/19GPpuFG4p9uG1sR_jVW8O3ptpoBfuQJxPuo5lIZf7hc/edit?usp=sharing

If there is anyone out there deeply familiar with NSF processes, and
has some time before October to help, that would be great. Not only
that I've tried to identify "organizational" things that someone with
more talent at that (if anyone knows anyone) that someone(s) could
solve as part of it. [0]

Basically the crux of my thoughts driving this was in testing hotel
and personal networks all over the USA over the last few months, and
gathering a few individuals also, on our mailing lists, to also do
flent testing, revealed that 99% of all wifi networks kind of stink in
some respect or another, and also, sadly, discovering the huge
deployment of essentially buggy implementations of fq_codel now
fielded as part of the survey I just completed for NLNET.[3]

As happy as I am with speedtest.net's new "working latency" app for
ios (try it!!!) it doesn't let a network engineer or driver writer
"get in there" and optimize their code for better latency like flent's
tests do, and most of the testing "out there" is reduced to single
number summaries of iperf3, it seems. I love flent, it's been a great
tool, I would like to find ways to get 100k more people using it
regularly without having to do a lot of extra work, and the methods
and analytics into things like wireshark, also. It needn't be this
flent foundation idea. Certainly I'm going to try to find a conference
or three in the coming years to talk to it - perhaps nanog, certainly
a wifi conference, etc. Perhaps others can pursue that, too.

I've never wanted to establish a non-profit for bufferbloat.net,
principally because I'd hoped we'd be done by now! I'd tried to
establish a more generic educational foundation (icei.org, if anyone
wants to see it on archive.org), which failed, tried for a floor under
cerowrt to keep it going, failed, tried to keep make wifi-fast going,
failed... (and the entire internet failed during that key
interview[]1) - it's always been easier to just "do the work" on no
money and on the backs of so many willing volunteers, than do anything
structured around it. I keep circulating a proposal for an "upgrade in
place", in the hope that someone grasps the seemingly viable business
models in it. And so it goes.

Basically with a string of failures like the streak I've long been on,
there is one person to point to for why it doesn't happen. Me.
Bothering to make the effort to find funding that we could use to more
effectively change the world than we have so far is now a once every
other year activity for me. I'm very glad that companies like canopus,
preseem, and others did find business models that worked. Me, I had
basically "quit" last year (jim is also mostly retired), and if it
wasn't for apple taking up the cause and so much progress elsewhere
this past year, I'd have stayed quit. As it is, I'm trying to nerve
myself up for doing something useful for - the summer at least - and I
don't know what that is.

(anyone going to "mountain connect" in colorado, may 23-25th?) .

Somewhere, out there, someone else must have a brilliant idea to
finally eradicate the bufferbloat from the world.

[0] I do wish I could find a concept with a profit model. But I
imagine the demand for test tools like this on android/ios/windows is
quite low, and the amount of investment required to produce one, quite
high. The mythological VC with a heart of gold, or billionaire with a
few extra bucks, would be more ideal.

[1] I'd really like to at the very least get merch for everyone here
that said something like:
    "I fixed the internet, and all I got was this lousy t-shirt".

[2] https://blog.cerowrt.org/post/bufferbloat_on_the_backbone/


-- 
[3] FQ World Domination pending:
https://blog.cerowrt.org/post/state_of_fq_codel/
Dave Täht CEO, TekLibre, LLC


More information about the Bloat mailing list