[Starlink] [Make-wifi-fast] Talk now up: How the internet really works

Dave Taht dave.taht at gmail.com
Fri Mar 11 10:51:59 EST 2022


On Thu, Mar 10, 2022 at 8:20 PM Vint Cerf <vint at google.com> wrote:
>
> just watching the first few minutes I could tell this was going to be a really fun talk - will watch the rest soon.


I cavort on the shoulders of giants!

I'm going to give myself an "A" for concept, but a "B" for
performance. The latter half could have gone a bit faster, and one of
the more difficult tricks, where one person already juggling tosses a
ball (new flow) to another to incorporate it into their flows, was too
hard to do, and didn't come off. Bunch of other flaws - notably I'd to
say over and over again it would be better to think about an optimum
for a low latency metaverse-capable internet would be for everyone to
think more about "steady kilobits per millisecond".

It's very hard to express how fair queuing works, also, correctly, in
the context of this talk. There are quite a few other networking
concepts that I hope could be explained in this way, the difficulties
with doing full duplex wireless using a water ballon to splatter the
reciever was originally part of the act but I cut it in deference to
the hotel staff!

> Dave, you should consider stand-up... :-)))

I've been thinking about retiring, and opening up a comedy club in
Starbase, Texas, with a marachi band on sundays.

After working on theory, code, standards, and serious publications for
so long, and being trapped behind zoom for the past 3 years, I needed
an live outlet, like this, to stretch out a bit. I miss the high speed
interactions you get out of "improv" for example, that's not a
zoom-able thing. After the conference Bruce invited me to a musical
jam which was *awesome* and also an example of what cannot presently
be done well over the present day internet. One day, I hope.

It was so great to get out and do this, and perhaps I'll do it again
one day or film it more carefully and not live.

If y'all would like a mostly serious explanation of many of the
problems wifi has (With a fun explanation towards the end of what all
OSes had been getting most wrong about wifi until then), please see
the 8 minute segment here:
https://www.youtube.com/watch?v=Rb-UnHDw02o&t=1550s or, preferably,
pass the whole thing to someone making wifi chips and drivers.

The related paper is rather dry in comparison, unless you get excited
about 10x reductions of network latency across the board in cdf plots.
https://www.usenix.org/conference/atc17/technical-sessions/presentation/hoilan-jorgesen
- a huge (and still flailing) goal for me has been to get the now
standard linux APIs for that into more wifi chipsets than just
Intel's, mediatek mt76, qualcomm ath9k and ath10k chips.

PS The only way I can think of to express how wireless signals degrade
over distance while using jugglers is via CGI, making the balls
diffuse and shrinking... Other ideas for how to express the inverse
square law simply, welcomed!

PPS (I'm quite curious as to how good the vanguard talk looked over
the much lower frame rate zoom participants, and how much freezing or
distortion of the feed they had)

> v
>
>
> On Thu, Mar 10, 2022 at 9:25 AM James Hurley <jamesghurley at gmail.com> wrote:
>>
>> Great talk Dave & MIT jugglers
>>
>> Dave is there any chance you can share the slides you were presenting in the video?
>>
>> > On 8 Mar 2022, at 16:07, Dave Taht <dave.taht at gmail.com> wrote:
>> >
>> > My talk last week at TTI Vanguard is now up on youtube. I used no
>> > slides, and a bunch of new simple analogies (including the world's
>> > most elaborate "rickroll"!) to explain the problems videoconferencing
>> > and voice have with competing with web traffic, with bits about
>> > cryptography, packet loss, bufferbloat, fair queueing and active queue
>> > management across all our access technologies today, as well as some
>> > notes as to the NTIA broadband programs and ongoing FCC measurements.
>> >
>> > https://www.youtube.com/watch?v=TWViGcBlnm0
>> >
>> > My special thanks to Len and Nancy Kleinrock, and to Lisa Yao, CEO of
>> > TTI Vanguard, for sponsoring my talk and travel. Plug: There were a
>> > bunch of very interesting other talks at that conference (including
>> > one amazing one on quantum computing that went over my head
>> > completely) that I enjoyed greatly. A huge thx also to my volunteers,
>> > Jamie, Vasu, and Joshua, who went the extra mile to help out.
>> >
>> > I am of course trying to reach new audiences outside our circle with
>> > my talk, please reshare widely?
>> >
>> > --
>> > I tried to build a better future, a few times:
>> > https://wayforward.archive.org/?site=https%3A%2F%2Fwww.icei.org
>> >
>> > Dave Täht CEO, TekLibre, LLC
>> > _______________________________________________
>> > Make-wifi-fast mailing list
>> > Make-wifi-fast at lists.bufferbloat.net
>> > https://lists.bufferbloat.net/listinfo/make-wifi-fast
>>
>> _______________________________________________
>> Starlink mailing list
>> Starlink at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/starlink
>
>
>
> --
> Please send any postal/overnight deliveries to:
> Vint Cerf
> 1435 Woodhurst Blvd
> McLean, VA 22102
> 703-448-0965
>
> until further notice
>
>
>


-- 
I tried to build a better future, a few times:
https://wayforward.archive.org/?site=https%3A%2F%2Fwww.icei.org

Dave Täht CEO, TekLibre, LLC


More information about the Starlink mailing list