You are a master of understatement in the below. I forwarded the
announcement around a bit.
Do we need to hire a PR flack to add excitement to future releases,
describing how flent can solve network problems, AND world hunger, AND
automate the production of dozens of academic papers?
Somewhat rel
http://www.brendangregg.com/blog/2016-11-30/linux-bcc-tcplife.html
also the bcc collection is interesting.
--
Dave Täht
Let's go make home routers and wifi faster! With better software!
http://blog.cerowrt.org
___
Flent-users mailing list
Flent-users@
Toke has been busy adding new features to the flent network test tool.
I consider it *almost* stable enough for a new release. Some of the
development has been focused on making the flent-gui much faster and
more responsive (as our data sets have got larger), others on
providing better default comm
I am cutting all the other lists off of this...
yes, I have had to patch netperf for more recent versions of gcc -
getting rid of an inline, as I recall. Haven't tried to build it
recently. I can go look.
I feel your pain re: RHEL and am glad you are chasing after fedora/etc.
On Tue, Dec 20, 201
gt;
>> That was what I needed to do for clang.
>>
>> -Aaron
>>
>> On Tue, Dec 20, 2016 at 18:16 Dave Taht wrote:
>>>
>>> I am cutting all the other lists off of this...
>>>
>>> yes, I have had to patch netperf for more recent vers
wrote:
> On Tue, Dec 20, 2016 at 11:02 AM, Dave Taht wrote:
>>
>> Active public servers include:
>>
>> flent-freemont.bufferbloat.net
>> ( this is colocated with flent-bbr-west which has bbr on by default - an
>> interesting test might be testing both these
https://www.vividcortex.com/blog/why-percentiles-dont-work-the-way-you-think
--
Dave Täht
Let's go make home routers and wifi faster! With better software!
http://blog.cerowrt.org
___
Flent-users mailing list
Flent-users@flent.org
http://flent.org/mail
this is not flent per se' blowing up, but your bloated path blowing
up, probably. :)
try a simpler test (tcp_upload or tcp_download) to start with, then
the nastier ones.
On Thu, Jan 5, 2017 at 7:21 AM, Jim Gettys wrote:
> Note that a simple ping during this run reports >42 seconds.
>
> I shou
On Thu, Jan 5, 2017 at 8:15 AM, Jim Gettys wrote:
>
> On Thu, Jan 5, 2017 at 11:12 AM, Dave Taht wrote:
>>
>> this is not flent per se' blowing up, but your bloated path blowing
>> up, probably. :)
>
>
> I ran ping myself during the test. The RTT's re
http://gafferongames.com/2016/07/21/launch-of-libyojimbo/
--
Dave Täht
Let's go make home routers and wifi faster! With better software!
http://blog.cerowrt.org
___
Flent-users mailing list
Flent-users@flent.org
http://flent.org/mailman/listinfo/flent-
https://github.com/iovisor/bcc
--
Dave Täht
Let's go make home routers and wifi faster! With better software!
http://blog.cerowrt.org
___
Flent-users mailing list
Flent-users@flent.org
http://flent.org/mailman/listinfo/flent-users_flent.org
Of all the codecs out there, I like opus best. (It's deeply in webrtc).
g711, and gsm, used to be the most common, but I've been out of this field
a long time.
___
Flent-users mailing list
Flent-users@flent.org
http://flent.org/mailman/listinfo/flent-use
On Mon, Nov 20, 2017 at 5:21 AM, Toke Høiland-Jørgensen <
notificati...@github.com> wrote:
> Pete Heist writes:
>
> >> On Nov 20, 2017, at 1:11 PM, Toke Høiland-Jørgensen <
> notificati...@github.com> wrote:
> >>
> >> Pete Heist writes:
> >>
> >> > G.711 can be simulated today with `-i 20ms -l 1
A goal for me has been to be able to run Opus at 24 bit, 96Khz, with 2.7ms
sampling latency.
Actually getting 8 channels of that through a loaded box would be mahvelous.
On Mon, Nov 20, 2017 at 1:14 PM, Dave Taht wrote:
>
>
> On Mon, Nov 20, 2017 at 5:21 AM, Toke Høiland-
good test. I wish I'd thought of it. Life below a ms is a PITA.
For comparison, an aarch64 and an x86_64:
root@nanopineo2:~# irtt sleep
Testing sleep accuracy...
Sleep DurationMean Error % Error
1ns 13.353µs 1335336.9
10ns 14.34µs
https://github.com/lucas-clemente/quic-go
--
Dave Täht
CEO, TekLibre, LLC
http://www.teklibre.com
Tel: 1-669-226-2619
___
Flent-users mailing list
Flent-users@flent.org
http://flent.org/mailman/listinfo/flent-users_flent.org
" Flent has been an invaluable network analysis tool while performing
these experiments and has become my goto application for testing
networking performance after making changes/optimizations/tweaks etc.
I have used it for testing both WAN performance as well as LAN
performance at 10Gbit across th
It's a bit flawed - not using https, nor a reasonable number of http
assets, averaging the http result rather than plotting it also, with
min/max... nor using flent, nor publishing their source widely so that
I could fiddle with setting up mininet to verify it produces correct
results...
but, it's
How widely available is netperf (netserver) for android? I know it's
in the ecosystem...
On Mon, Oct 8, 2018 at 8:32 AM Toke Høiland-Jørgensen wrote:
>
> Louie Lu writes:
>
> > Not now, it is in my plan to update
> > , the goal is to make this app act like magiciperf*, which accept command
> > li
Linode is making available nanoservers for 5/month. (was paying
20/month two years ago)...
A) flent-singapore.taht.net. I would actually like flent.org to seize
control of all
these domains...
B) linode's kernel ships and makes available bbr by default now. It's not even
a kernel module.
C) on th
and, oh, yea, fq_codel is the default qdisc. I really wonder how much
of the world is now running it.
___
Flent-users mailing list
Flent-users@flent.org
http://flent.org/mailman/listinfo/flent-users_flent.org
for it.
how much app is needed to wrap for:
system("netserver -N");
? seriously. I have no idea.
we've had anyone with any chops at android... and I also confess being curious
if the babel 1.8.3 daemon worked on android?
>
> Thanks,
> Louie.
> Dave Taht 於 2018年10月9日
so I just upgraded to ubuntu 18.04 finally and tried to get flent running.
flent's in toke's ppa... installs irtt and flent...
no netperf in their default repos
the multiverse/universe version is 2.6.
netperf 2.7 won't build due to a missing makeinfo utility that's also
in the universe repo
I just turned it on again (I think it had crashed).
I would like to enable irtt fully across the flent network, and with
liberal defaults (longer tests, shorter intervals). Can you suggest a
useful set of defaults to use?
Another option is to make several versions available on different port
numb
I guess so... send me a ssh key?
On Fri, Jan 25, 2019 at 12:26 AM Pete Heist wrote:
>
>
> > On Jan 11, 2019, at 5:31 PM, Pete Heist wrote:
> >
> >> On Jan 11, 2019, at 5:00 PM, Dave Taht wrote:
> >>
> >> I just turned it on again (I think it had c
My principal desires for udp-lite support in both netperf and irtt is
that it is a udp alternative that actually works, at least on linux
and some forms of bsd, over ipv6, and so far as I can tell (without
serious testing) ipv4, even through nat.
UDP port space is especially constrained in light o
I note your "measure corruption" thing is a cool idea, I'd not thunk
of that too!
___
Flent-users mailing list
Flent-users@flent.org
http://flent.org/mailman/listinfo/flent-users_flent.org
I see from the iccrg preso at 7 minutes 55 s in, that there is a test
described as:
20 BBRv2 flows
starting each 100ms, 1G, 1ms
Linux codel with ECN ce_threshold at 242us sojurn time.
I interpret this as
20 flows, starting 100ms apart
on a 1G link
with a 1ms transit time
and linux codel with ce_
Thanks!
On Fri, Apr 5, 2019 at 5:11 PM Neal Cardwell wrote:
>
> On Fri, Apr 5, 2019 at 3:42 AM Dave Taht wrote:
>>
>> I see from the iccrg preso at 7 minutes 55 s in, that there is a test
>> described as:
>>
>> 20 BBRv2 flows
>> starting each 100ms, 1G
> With fq_codel and the same ECN marking threshold (fq_codel ce_threshold
> 242us), we see slightly smoother fairness properties (not surprising) but
> with slightly higher latency.
>
> The basic summary:
>
> retransmits: 0
> flow throughput: [46.77 .. 51.48]
> RTT samples at various percentiles:
this commit from the immortal edumazet arrived this morning in net-next:
https://www.spinics.net/lists/netdev/msg577452.html
...
> Adding delays to TCP flows is crucial for studying behavior
> of TCP stacks, including congestion control modules.
>
> Linux offers netem module, but it has unpracti
The new linode data center finally went up last week in mumbai. Aside
from the ones in singapore and japan,
this is the furthest "out" of any flent servers we have.
sce-mumbai.taht.net # name might change
Right now this server is just under test!! it's the first kernel I've
had up with spectre mi
rrul, not rrule
On Fri, Jan 24, 2020 at 12:03 PM Stephen Hemminger
wrote:
>
> Am trying to run flent in a network namespace and it won't work.
> Seems to demand connection to the Internet for DNS.
>
> The scenario is run net server in one namespace, and flent in another.
> Forward packets out an
sudo python3 setup.py install
or
make install
flent is actually measuring rate differences, it does not actually assign
bandwidth. So your result - where the first flow grabs more bandwidth than
the others, is an artifact of your queuing (probably fifo and tcp cubic).
The first flow wins. In the c
Just because I'm dumb, why are you fiddling with non-standard dscp markings
in the first place?
I've been trying to get the folk testing this set for a while now,
https://tools.ietf.org/html/draft-ietf-tsvwg-rtcweb-qos-18
they are essentially supported in sch_cake's 4 tier model.
On Tue, Apr 2
good to know those markings, thx. So far as I recall riverbed ships cake.
Can you confirm?
It's easy to modify the prio table but feedback as to what is in use
elsewhere has always been lacking.
On Tue, Apr 28, 2020 at 10:41 AM olg33 wrote:
> Right, using standard markings would make everyone
On Tue, Apr 28, 2020 at 1:03 PM olg33 wrote:
> Ohh, right, the markings_map is only used for irtt and iperf, for
> netperf the values are just passed directly through. Totally forgot
> about that. I guess I could fix that, but TBH I'm not terribly keen on
> supporting non-standard diffserv markin
https://forum.openwrt.org/t/fq-codel-fast-helpers-test-and-testers/63037/60
--
"For a successful technology, reality must take precedence over public
relations, for Mother Nature cannot be fooled" - Richard Feynman
d...@taht.net CTO, TekLibre, LLC Tel: 1-831-435-0729
__
See: https://arxiv.org/pdf/2012.14996.pdf
Things I really like:
* they used flent
* Using "variance" as the principal signal. This is essentially one of
the great unpublished and unanalyzed improvements on the minstrel
algorithm as well
* Conventional ecn response
* outperforms bbr on variable li
I am hoping to finally get some time to participate in revising the
wikipedia articles soon.
I've also been dusting off old pieces that could use revision.
I'd actually thought I'd published this one:
https://github.com/dtaht/blog-cerowrt/blob/master/content/post/found_in_flent.md
But didn't. Y
Currently centos (and I assume redhat) is at 4.18. Cake went into 4.19
so I assume the next major
redhat/centos releases will have it.
Is there a yum/rpm expert in the house? flent does not appear to be
packaged up for this (?),
neither is netperf or irtt. Is there a repo I could use?
tc is not s
"ERROR: Unable to find a usable Qt version."
On Wed, Mar 9, 2022 at 2:08 PM Nathan Owens wrote:
>
> Yea, I tried all that, and followed the directions to install on my mac
> freshly, no luck!
>
> On Wed, Mar 9, 2022 at 11:05 AM Dave Taht wrote:
>>
>> a
if anyone has a bit of spare time this holiday, I'd like to collect a
few results from various networks around the web this week, not just
on the turris. Example command line and outputs here:
https://forum.turris.cz/t/sqm-on-turris-flent-benchmarks/17048/
--
I tried to build a better future, a
fq-quantum 300
> limit 10240
> rate 8mbit
> }
> upload {
> ecn enable
> flows 1024
> fq-quantum 300
> limit 10240
> rate 600kbit
> target 15ms
> }
that
they make backlog, packet and ,drop statistics easily available across
their OSes. I'm pretty sure that at higher bandwidths their default
settings for SFQ and pfifo are too low...
>
> The OneDrive link to my flent files should still be active in that post as
> well.
>
>
I am in the middle of writing a modest proposal for the NSF, to put
more of an organisational floor under it, and am wondering if there
are any other flent users out there, willing to submit a "letter of
collaboration" as part of it?
I'm also kind of curious how many hits per month flent.org gets?
ending the money on good stuff,
which I'm *really good at*. But certainly I agree with the overall
thrust being to get some sort of administrative help for the things we
don't do well.
On Sun, May 8, 2022 at 8:23 AM Toke Høiland-Jørgensen wrote:
>
> Dave Taht writes:
>
> >
Just filed this feature request over here:
https://github.com/network-quality/goresponsiveness/issues/19
--
FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
Dave Täht CEO, TekLibre, LLC
___
Flent-users mailing list -- flent
OT... Those latency numbers are spectacular. Any hint as to why? :)
On Tue, Aug 9, 2022, 12:48 PM Toke Høiland-Jørgensen wrote:
> Erik Taraldsen writes:
>
> > I'm trying to automate testing and graphing with flent. For my purposes
> it
> > probably will make sense to use CDF and box plots. Ho
https://github.com/m-lab/tcp-info/
--
FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
Dave Täht CEO, TekLibre, LLC
___
Flent-users mailing list -- flent-users@flent.org
To unsubscribe send an email to flent-users-le...@flen
Hopefully we mostly fixed the mt76 chips recently,
https://www.linkedin.com/feed/update/urn:li:activity:6973288514096414720/
ax is still a pita
On Tue, Aug 9, 2022 at 10:59 PM Erik Taraldsen wrote:
>
>
> On Tue, Aug 9, 2022 at 10:01 PM Dave Taht wrote:
>>
>> OT... Thos
I've been working on a script that exercises some old tests. My goal
was to get something that ran
in under an hour that got a comprehensive viewpoint.
I'm terrible at python, so I do things in shell, rather than the batch facility.
Anyway, presently several of the more obscure tests break in
thi
seeing bbr not get out of slow start, then get in a fight with cubic,
is pretty hilarious...
--
FQ World Domination pending: https://blog.cerowrt.org/post/state_of_fq_codel/
Dave Täht CEO, TekLibre, LLC
___
Flent-users mailing list -- flent-users@flent.
https://ieeexplore.ieee.org/abstract/document/9887655
--
This song goes out to all the folk that thought Stadia would work:
https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-698135607352320-FXtz
Dave Täht CEO, TekLibre, LLC
___
Flent-u
I really enjoyed jim roskin's talk below, and especially the
animations of histogram progress over time.
https://www.youtube.com/watch?v=_uaaCiyJCFA&t=1120s
I have largely relied on static cdf plots to accomplish the same
thing, but I can'd help but wonder what it would take to produce
an animati
A rather long, meandering piece:
https://blog.cerowrt.org/post/flaws_in_flent/
--
This song goes out to all the folk that thought Stadia would work:
https://www.linkedin.com/posts/dtaht_the-mushroom-song-activity-698135607352320-FXtz
Dave Täht CEO, TekLibre, LLC
_
On Sun, Jan 15, 2023 at 11:14 PM Sebastian Moeller wrote:
>
> Hi Dave,
>
> On 16 January 2023 02:52:45 CET, Dave Taht wrote:
> >A rather long, meandering piece:
> >
> >https://blog.cerowrt.org/post/flaws_in_flent/
>
> I think 'Flent’s default sampling r
nd I have long regarded "3rd world" and
"less developed", as insulting, also.
I have no idea how well george carlin translates into other cultures,
but he is a hero of mine, as is the BOFH.
https://www.youtube.com/watch?v=o25I2fzFGoY&themeRefresh=1
>
> Regards
>
58 matches
Mail list logo
