[Make-wifi-fast] An "Accurate" Platform for Investigating TCP Performance in Wi-Fi Networks

[Make-wifi-fast] An "Accurate" Platform for Investigating TCP Performance in Wi-Fi Networks

[Dave Taht]

It is hard to contain my disappointment that a modern work on wifi,
that even mentions bufferbloat, could have missed our work on fixing
wifi (such as "ending the anomaly", now the default APIs in the linux
kernel), and make claims for the emulator that are so completely
wrong, and further claim that the behaviors matched a real world AP.

re: https://mdpi-res.com/d_attachment/futureinternet/futureinternet-15-00246/article_deploy/futureinternet-15-00246.pdf?version=1689771540

Can I suggest strongly to the authors that they try to understand rate
vs range vs buffering, the intersection of the half duplex txop
infrastructure with multiple stations, the side effects of
aggregation, and other scheduling difficulties mininet does not come
even close to handling?

20 packet buffers?? A more typical number is much greater.
Test through walls or at distance?

Here are 4 resources that might help deepen your understanding.

https://www.youtube.com/watch?v=Rb-UnHDw02o&t=1532s (10 minutes here,
but do the whole thing)
https://bufferbloat-and-beyond.net/
https://www.cs.kau.se/tohojo/airtime-fairness/
https://blog.cerowrt.org/post/minstrel/

-- 
Podcast: https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
Dave Täht CSO, LibreQos

Re: [Make-wifi-fast] An "Accurate" Platform for Investigating TCP Performance in Wi-Fi Networks

[Bob McMahon]

[-- Attachment #1.1: Type: text/plain, Size: 3228 bytes --]

I think making WiFi tech accessible to students is a very good thing. That
students make mistakes is also a good thing. I understand the frustration
though.

I've recently rolled out a raspberry pi 4 crib sheet
<https://www.l4sgear.com/> on how to build an L4S capable device. It's a
fun project and comes in at around $100 U.S.  Throw in some switches, USB
WiFi cards, ethernet for control, a GPS hat for pulse per second, use iperf
2 and pyflows. All for less than $500.

The expensive part is placing the RF under control
<https://vaunix.com/digital-phase-shifters/> but that's doable too - though
likely in the few thousands of dollars and maybe just a bit more. The math
is already done thanks to those studying molecular evolution
<https://youtu.be/PNoUcQTCxiM>.

Bob

On Wed, Jul 19, 2023 at 5:30 PM Dave Taht via Make-wifi-fast <
make-wifi-fast@lists.bufferbloat.net> wrote:

> It is hard to contain my disappointment that a modern work on wifi,
> that even mentions bufferbloat, could have missed our work on fixing
> wifi (such as "ending the anomaly", now the default APIs in the linux
> kernel), and make claims for the emulator that are so completely
> wrong, and further claim that the behaviors matched a real world AP.
>
> re:
> https://mdpi-res.com/d_attachment/futureinternet/futureinternet-15-00246/article_deploy/futureinternet-15-00246.pdf?version=1689771540
>
> Can I suggest strongly to the authors that they try to understand rate
> vs range vs buffering, the intersection of the half duplex txop
> infrastructure with multiple stations, the side effects of
> aggregation, and other scheduling difficulties mininet does not come
> even close to handling?
>
> 20 packet buffers?? A more typical number is much greater.
> Test through walls or at distance?
>
> Here are 4 resources that might help deepen your understanding.
>
> https://www.youtube.com/watch?v=Rb-UnHDw02o&t=1532s (10 minutes here,
> but do the whole thing)
> https://bufferbloat-and-beyond.net/
> https://www.cs.kau.se/tohojo/airtime-fairness/
> https://blog.cerowrt.org/post/minstrel/
>
> --
> Podcast:
> https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
> Dave Täht CSO, LibreQos
> _______________________________________________
> Make-wifi-fast mailing list
> Make-wifi-fast@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/make-wifi-fast

-- 
This electronic communication and the information and any files transmitted 
with it, or attached to it, are confidential and are intended solely for 
the use of the individual or entity to whom it is addressed and may contain 
information that is confidential, legally privileged, protected by privacy 
laws, or otherwise restricted from disclosure to anyone else. If you are 
not the intended recipient or the person responsible for delivering the 
e-mail to the intended recipient, you are hereby notified that any use, 
copying, distributing, dissemination, forwarding, printing, or copying of 
this e-mail is strictly prohibited. If you received this e-mail in error, 
please return the e-mail to the sender, delete it from your computer, and 
destroy any printed copy of it.

[-- Attachment #1.2: Type: text/html, Size: 4548 bytes --]

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]

Re: [Make-wifi-fast] An "Accurate" Platform for Investigating TCP Performance in Wi-Fi Networks

[Dave Taht]

On Wed, Jul 19, 2023 at 6:48 PM Bob McMahon <bob.mcmahon@broadcom.com> wrote:
>
> I think making WiFi tech accessible to students is a very good thing. That students make mistakes is also a good thing.

https://github.com/open-sdr/openwifi is making half decent progress.

>I understand the frustration though.

I had opened that paper hoping for something more.

>
> I've recently rolled out a raspberry pi 4 crib sheet on how to build an L4S capable device. It's a fun project and comes in at around $100 U.S.  Throw in some switches, USB WiFi cards, ethernet for control, a GPS hat for pulse per second, use iperf 2 and pyflows. All for less than $500.
>
> The expensive part is placing the RF under control but that's doable too - though likely in the few thousands of dollars and maybe just a bit more. The math is already done thanks to those studying molecular evolution.

That FPGA above looks promising.

>
> Bob
>
> On Wed, Jul 19, 2023 at 5:30 PM Dave Taht via Make-wifi-fast <make-wifi-fast@lists.bufferbloat.net> wrote:
>>
>> It is hard to contain my disappointment that a modern work on wifi,
>> that even mentions bufferbloat, could have missed our work on fixing
>> wifi (such as "ending the anomaly", now the default APIs in the linux
>> kernel), and make claims for the emulator that are so completely
>> wrong, and further claim that the behaviors matched a real world AP.
>>
>> re: https://mdpi-res.com/d_attachment/futureinternet/futureinternet-15-00246/article_deploy/futureinternet-15-00246.pdf?version=1689771540
>>
>> Can I suggest strongly to the authors that they try to understand rate
>> vs range vs buffering, the intersection of the half duplex txop
>> infrastructure with multiple stations, the side effects of
>> aggregation, and other scheduling difficulties mininet does not come
>> even close to handling?
>>
>> 20 packet buffers?? A more typical number is much greater.
>> Test through walls or at distance?
>>
>> Here are 4 resources that might help deepen your understanding.
>>
>> https://www.youtube.com/watch?v=Rb-UnHDw02o&t=1532s (10 minutes here,
>> but do the whole thing)
>> https://bufferbloat-and-beyond.net/
>> https://www.cs.kau.se/tohojo/airtime-fairness/
>> https://blog.cerowrt.org/post/minstrel/
>>
>> --
>> Podcast: https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
>> Dave Täht CSO, LibreQos
>> _______________________________________________
>> Make-wifi-fast mailing list
>> Make-wifi-fast@lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/make-wifi-fast
>
>
> This electronic communication and the information and any files transmitted with it, or attached to it, are confidential and are intended solely for the use of the individual or entity to whom it is addressed and may contain information that is confidential, legally privileged, protected by privacy laws, or otherwise restricted from disclosure to anyone else. If you are not the intended recipient or the person responsible for delivering the e-mail to the intended recipient, you are hereby notified that any use, copying, distributing, dissemination, forwarding, printing, or copying of this e-mail is strictly prohibited. If you received this e-mail in error, please return the e-mail to the sender, delete it from your computer, and destroy any printed copy of it.



-- 
Podcast: https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
Dave Täht CSO, LibreQos

Re: [Make-wifi-fast] An "Accurate" Platform for Investigating TCP Performance in Wi-Fi Networks

[Bob McMahon]

[-- Attachment #1.1: Type: text/plain, Size: 5064 bytes --]

yeah, hoping for more can for sure cause frustration when things aren't
where we'd like them to be or where we think they should be.  I was told
early in my career by our company leadership, "See the world as it is, not
as you want it to be."  I never really liked that so I modified it a bit
to, "See the world as it is and help towards making it what you know it can
become." Then we march along on that journey and never finish so
frustration will be our companion at times ;)

Bob

On Wed, Jul 19, 2023 at 6:55 PM Dave Taht <dave.taht@gmail.com> wrote:

> On Wed, Jul 19, 2023 at 6:48 PM Bob McMahon <bob.mcmahon@broadcom.com>
> wrote:
> >
> > I think making WiFi tech accessible to students is a very good thing.
> That students make mistakes is also a good thing.
>
> https://github.com/open-sdr/openwifi is making half decent progress.
>
> >I understand the frustration though.
>
> I had opened that paper hoping for something more.
>
> >
> > I've recently rolled out a raspberry pi 4 crib sheet on how to build an
> L4S capable device. It's a fun project and comes in at around $100 U.S.
> Throw in some switches, USB WiFi cards, ethernet for control, a GPS hat for
> pulse per second, use iperf 2 and pyflows. All for less than $500.
> >
> > The expensive part is placing the RF under control but that's doable too
> - though likely in the few thousands of dollars and maybe just a bit more.
> The math is already done thanks to those studying molecular evolution.
>
> That FPGA above looks promising.
>
> >
> > Bob
> >
> > On Wed, Jul 19, 2023 at 5:30 PM Dave Taht via Make-wifi-fast <
> make-wifi-fast@lists.bufferbloat.net> wrote:
> >>
> >> It is hard to contain my disappointment that a modern work on wifi,
> >> that even mentions bufferbloat, could have missed our work on fixing
> >> wifi (such as "ending the anomaly", now the default APIs in the linux
> >> kernel), and make claims for the emulator that are so completely
> >> wrong, and further claim that the behaviors matched a real world AP.
> >>
> >> re:
> https://mdpi-res.com/d_attachment/futureinternet/futureinternet-15-00246/article_deploy/futureinternet-15-00246.pdf?version=1689771540
> >>
> >> Can I suggest strongly to the authors that they try to understand rate
> >> vs range vs buffering, the intersection of the half duplex txop
> >> infrastructure with multiple stations, the side effects of
> >> aggregation, and other scheduling difficulties mininet does not come
> >> even close to handling?
> >>
> >> 20 packet buffers?? A more typical number is much greater.
> >> Test through walls or at distance?
> >>
> >> Here are 4 resources that might help deepen your understanding.
> >>
> >> https://www.youtube.com/watch?v=Rb-UnHDw02o&t=1532s (10 minutes here,
> >> but do the whole thing)
> >> https://bufferbloat-and-beyond.net/
> >> https://www.cs.kau.se/tohojo/airtime-fairness/
> >> https://blog.cerowrt.org/post/minstrel/
> >>
> >> --
> >> Podcast:
> https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
> >> Dave Täht CSO, LibreQos
> >> _______________________________________________
> >> Make-wifi-fast mailing list
> >> Make-wifi-fast@lists.bufferbloat.net
> >> https://lists.bufferbloat.net/listinfo/make-wifi-fast
> >
> >
> > This electronic communication and the information and any files
> transmitted with it, or attached to it, are confidential and are intended
> solely for the use of the individual or entity to whom it is addressed and
> may contain information that is confidential, legally privileged, protected
> by privacy laws, or otherwise restricted from disclosure to anyone else. If
> you are not the intended recipient or the person responsible for delivering
> the e-mail to the intended recipient, you are hereby notified that any use,
> copying, distributing, dissemination, forwarding, printing, or copying of
> this e-mail is strictly prohibited. If you received this e-mail in error,
> please return the e-mail to the sender, delete it from your computer, and
> destroy any printed copy of it.
>
>
>
> --
> Podcast:
> https://www.linkedin.com/feed/update/urn:li:activity:7058793910227111937/
> Dave Täht CSO, LibreQos
>

-- 
This electronic communication and the information and any files transmitted 
with it, or attached to it, are confidential and are intended solely for 
the use of the individual or entity to whom it is addressed and may contain 
information that is confidential, legally privileged, protected by privacy 
laws, or otherwise restricted from disclosure to anyone else. If you are 
not the intended recipient or the person responsible for delivering the 
e-mail to the intended recipient, you are hereby notified that any use, 
copying, distributing, dissemination, forwarding, printing, or copying of 
this e-mail is strictly prohibited. If you received this e-mail in error, 
please return the e-mail to the sender, delete it from your computer, and 
destroy any printed copy of it.

[-- Attachment #1.2: Type: text/html, Size: 6977 bytes --]

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4206 bytes --]