Re: [Cerowrt-devel] [Make-wifi-fast] perverse powersave bug with sta/ap mode

Has anyone modeled what the multicast to multiple-unicast efficiency threshold is? The point where you go from it being more efficient to send multicast traffic to individual STAs instead of sending a monstrous (in time) multicast-rate packet? 2, 5, 10 STAs? The per-STA-queue work should make th

Re: [Cerowrt-devel] using a usb3 hub as a hackerboard power switch?

Most USB hub chips I've seen don't have individual lines per-port. Instead they monitor the +5VDC rail for powering all USB devices, and grant requests to use power based on a configured power budget (how many amps the supply can push). To hack it, I'd use a Pi with it's relay modules and cut the

Re: [Cerowrt-devel] bugs using the 1200ac as a target

That's behavior I've seen multiple times in the past, but I never had a good feel for what the cause was. -Aaron On Wed, Apr 27, 2016 at 3:08 PM, Dave Taht wrote: > run netserver on the router. It doesn't seem to matter what qdisc is > on the router. > > hit it with the flent rrul_be or rrul te

Re: [Cerowrt-devel] [Make-wifi-fast] what is the right way to do, and rationale, for wifi channel scans?

wpa_supplicant? I've seen some atrocious behavior from wpa_supplicant in the past (especially the closed-source versions from some chipset companies). One reason it's likely to be scanning is that it's looking for alternate APs for the same SSID, which are "better". I saw one implementation that

Re: [Cerowrt-devel] [Make-wifi-fast] wifi airtime fairness patches could use eyeballs and testing

On Sun, Aug 14, 2016 at 3:21 PM, Dave Taht wrote: > > Ideas as to why crypto chops off so much throughput in the first place > (from 120mbit to 80), and as to what else to poke into with crypto on, > would be nice. Crypto as in WPA2? If 1 client is fine, but multiple clients is not, my guess i

Re: [Cerowrt-devel] [Make-wifi-fast] wifi airtime fairness patches could use eyeballs and testing

On Sun, Aug 14, 2016 at 4:02 PM, Dave Taht wrote: > On Mon, Aug 15, 2016 at 12:49 AM, Aaron Wood wrote: > > On Sun, Aug 14, 2016 at 3:21 PM, Dave Taht wrote: > >> > >> > >> Ideas as to why crypto chops off so much throughput in the first place > >>

Re: [Cerowrt-devel] BBR congestion control algorithm for TCP in net-next

Dumb question on this: The tcp_bbr_info struct for a socket can be inspected at runtime through the ss utility or through a get socket opts call, right? -Aaron On Sat, Sep 17, 2016 at 11:34 AM, Maciej Soltysiak wrote: > Hi, > > Just saw this: https://patchwork.ozlabs.org/patch/671069/ > > Inte

Re: [Cerowrt-devel] BBR congestion control algorithm for TCP in net-next

Thanks! And sorry that I missed the sample code in the patch. On Mon, Sep 26, 2016 at 12:30 Neal Cardwell wrote: > On Mon, Sep 26, 2016 at 2:47 PM, Aaron Wood wrote: > > Dumb question on this: The tcp_bbr_info struct for a socket can be > > inspected at runtime through th

Re: [Cerowrt-devel] BBR congestion control algorithm for TCP in net-next

bloat.net is propagating (both IPv4 and IPv6) > > Til then: > the ipv4 for the thing is: 173 dot 230 dot 156 dot 252 > > I've put a broad sweep of results up in my github, haven't looked at > them to a huge extent yet. > > > On Mon, Sep 26, 2016 at 12:45 PM, Aaro

Re: [Cerowrt-devel] [Bloat] Comcast's NANOG slides re Bufferbloat posted (Oct 2016)

I need to box my test unit up and return it, but my area has 160/12 service if you get the upgraded rates (which I do) -Aaron On Thu, Oct 20, 2016 at 11:48 Klatsky, Carl wrote: > On Thu, 20 Oct 2016, Rich Brown wrote: > > > https://www.nanog.org/sites/default/files/20160922_Klatsky_First_Steps >

Re: [Cerowrt-devel] WNDR3800 improvements?

I thought it did better than that with Cake? -Aaron On Tue, Dec 13, 2016 at 8:41 PM, Dave Täht wrote: > > > On 12/13/16 4:58 PM, Matt Taggart wrote: > > Hi, > > > > I love the WNDR3800 platform, it's been great over the years first with > > cerowrt and then openwrt. Of the many I've deployed I

Re: [Cerowrt-devel] [Bloat] flent testers wanted prior to next release

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 servers at the same time > via the rtt_fair* tests from

Re: [Cerowrt-devel] flent testers wanted prior to next release

On Tue, Dec 20, 2016 at 12:20 PM, Joel Wirāmu Pauling wrote: > My biggest bug bear is that reliance on netperf/netserver with -DEMO mode > compilation time flag breaks compilation on recent RHEL and Fedora boxes > due to recent GCC incompatibilities. > I ran into some issues on OSX due to the gc

Re: [Cerowrt-devel] gaming routers...

"but it has revealed that the WRT32X is designed to automatically detect computers using the Killer-line of network adapters , indicating it's probably a gaming-focused PC from a company like Alienware, MSI, or Razer that requires priority access to the home's inte

Re: [Cerowrt-devel] gaming routers...

quot; QoS handling. -Aaron On Wed, Jan 4, 2017 at 13:57 Jonathan Morton wrote: > > > On 4 Jan, 2017, at 23:45, Aaron Wood wrote: > > > > "but it has revealed that the WRT32X is designed to automatically detect > computers using the Killer-line of network adapters, in

Re: [Cerowrt-devel] fixing reboots behind comcast modems

I've not had to reboot my modem (SB6141) when rebooting my WRT1900AC (running stock fw, OpenWRT, or LEDE). I always have to reboot the SB if I change the router out (different MAC). I've had the same experiences with all the Surfboard-line modems I've had in the past (going back to 2004). -Aaron

Re: [Cerowrt-devel] edgerouter Xes and lede

> I am currently using ac1750's as my air interfaces and hanging them off > Ethernet over powerline or directly cabled via UTP and bringing all the > routing back to x86 based vm's for l3. > I would be wary of the ethernet over powerline units, they have been massively (100s of ms) overbloated in

Re: [Cerowrt-devel] bcp38 and the caida "spoofer" tool

I usually have it running, but I just checked and see that it didn't get installed the last time I setup my router. So thanks for the reminder to get it re-installed and re-configured (with the IP of my cable-modem white-listed). -Aaron On Mon, Apr 17, 2017 at 9:31 AM, Dave Taht wrote: > > fro

Re: [Cerowrt-devel] solar wifi ap designs?

I'd pick an AP that works off of a 12vdc supply, and then try to setup a 12vdc solar supply (which should get you lots of stuff for automotive/rv use). Are you planning on deep cycle lead acid or lithium ion for storage? -Aaron On Sun, Jun 4, 2017 at 17:49 Dave Taht wrote: > I keep finding nice

Re: [Cerowrt-devel] solar wifi ap designs?

wrote: > lithium ion... although I admit a fondness for edison batteries. > > On Sun, Jun 4, 2017 at 6:54 PM, Aaron Wood wrote: > > I'd pick an AP that works off of a 12vdc supply, and then try to setup a > > 12vdc solar supply (which should get you lots of stuff

Re: [Cerowrt-devel] solar wifi ap designs?

As for why 12V: it's a common standard regulator that allows all the various internal voltages to be derived from (5, 3.3, 1.8, etc). So that's why a lot of designs use that. Especially since the 5v for USB can be isolated corrected against being pulled too low by devices that are plugged in. That

Re: [Cerowrt-devel] anyone fiddlng with these?

"securely sending messages to the web" -> "Sending telemetry data to my cloud-based data processing pipeline". Both MQTT (over TLS) and HTTPS are both used heavily for sending data upstream. Some companies in this space are thinking about security, others... less so. In general, the cloud provid

Re: [Cerowrt-devel] fq_pie for linux

On Wed, Dec 5, 2018 at 11:51 PM Toke Høiland-Jørgensen wrote: > Dave Taht writes: > > > https://github.com/gautamramk/FQ-PIE-for-Linux-Kernel/issues/2 > > With all the variants of fq+AQM, maybe decoupling the FQ part and the > AQM part would be worthwhile, instead of reimplementing it for each >

Re: [Cerowrt-devel] dlte

On Tue, Dec 11, 2018 at 11:02 Dave Taht > > Well, I kind of envision that happening, unless we make wifi work really > really well again. I loved learning about low-e glass recently, I think > that is going to interfere with celluar, and make wifi deployments better. It definitely messes with ce

Re: [Cerowrt-devel] I so love seeing stuff like this

What issues do you have with the wifi on the WRT1900AC? Mine works great. Although it struggles with cake at >150Mbps or so (when also routing IPv4/NAT). But then I'm still on 17.01.1 with driver v10.3.2.0-20170110 On Fri, Feb 1, 2019 at 12:43 PM David Lang wrote: > On Fri, 1 Feb 2019, Mikael

Re: [Cerowrt-devel] I so love seeing stuff like this

Yes, I have the v1 (mamba). I jumped on it when it came out to get a multicore arm with 802.11ac On Sat, Feb 2, 2019 at 11:56 PM Mikael Abrahamsson wrote: > On Sat, 2 Feb 2019, Aaron Wood wrote: > > > What issues do you have with the wifi on the WRT1900AC? Mine works > >

Re: [Cerowrt-devel] [Bloat] some sqm and cpu performance numbers for a variety of devices

40 AM Dave Taht wrote: > > https://forum.openwrt.org/t/comparative-throughput-testing-including-nat-sqm-wireguard-and-openvpn/44724/44 > > (H/T to aaron wood) > > The post persistently points out that openvpn tends to optimize for > one direction only. This is in part due to th

Re: [Cerowrt-devel] openwrt freeze jan 1

Wait, the AQL changes for ath10k aren't in 19.07? I'm seeing the aqm/airtime fairness controls/stats, but don't know what to look for in /sys/kernel to see if AQL is present (if it's visible there at all), or if I'm going to have to run side-by-side tests... Relatedly, the current 80211 queue set

Re: [Cerowrt-devel] semi off-topic: LEDs on WRT1200AC

As they're white LEDs, voltage (well, the current-limiting resistor used) will probably matter (different color LEDs tend to take different voltages/currents for the same brightness) On Tue, Mar 16, 2021 at 8:06 PM Jonathan Morton wrote: > > On 17 Mar, 2021, at 3:24 am, Valdis Klētnieks > wrote

Re: [Cerowrt-devel] Looking for MORE SQM Router Recommendations !

I'm using the Seeed Odyssey https://www.seeedstudio.com/ODYSSEY-X86J4105864-p-4447.html It will shape inbound using cake (piece of cake) at 1Gbps. Dual Intel i210 controllers. I'm running stable OpenWRT, although I needed to use ToT to get a UEFI boot setup for it that would work with the intern

Re: [Cerowrt-devel] [Bloat] Little's Law mea culpa, but not invalidating my main point

On Mon, Jul 12, 2021 at 1:32 PM Ben Greear wrote: > UDP is better for getting actual packet latency, for sure. TCP is > typical-user-experience-latency though, > so it is also useful. > > I'm interested in the test and visualization side of this. If there were > a way to give engineers > a good

Re: [Cerowrt-devel] [Make-wifi-fast] [Bloat] Little's Law mea culpa, but not invalidating my main point

With the disclaimer that I'm not as strong in statistics and modelling as I'd like to be I think it's not useful to attempt to stochastically model the behavior of what are actually active (well, reactive) components. The responses of each piece are deterministic, but the inputs (users) are n

Re: [Cerowrt-devel] [Make-wifi-fast] [Starlink] [Cake] Due Aug 2: Internet Quality workshop CFP for the internet architecture board

My own experiments with this, in the past (5+ years ago), was that you absolutely had to use cabled setups for repeatability, but then didn't have enough randomness in the variability to really test anything that was problematic. We could create hidden nodes, or arbitrary meshes of devices, but th

Re: [Cerowrt-devel] 10gige and 2.5gige

On Sun, Dec 19, 2021 at 10:07 AM David P. Reed wrote: > I know I shouldn't kink or pull fiber hard. In the worst case, I pull > light flexible conduit through walls with pull strings so I can add > arbitrary numbers of fibers. This is good practice, anyway (for wires or > fibers). > This is what

Re: [Cerowrt-devel] ML

Only if you tell it to... For stuff like this, my understanding is that the hard part is always in the labelling of the training data. I've been ML-adjacent for the last ~decade, but not doing ML myself. I really should sit down and do some model training on various kinds of datasets that would

Re: [Cerowrt-devel] new beaglebone

At least for the sub-1GHz 802.15.4, that’s for range that wifi can’t get, due to its lower signal loss over distance. But 10Mbut Ethernet’s weird (to me) -Aaron On Wed, Mar 8, 2023 at 6:45 AM Dave Taht via Cerowrt-devel < cerowrt-devel@lists.bufferbloat.net> wrote: > Modestly weird to see 10Mbi

Re: [Cerowrt-devel] new beaglebone

-bus model that ControlNet uses. -Aaron On Thu, Mar 9, 2023 at 8:18 AM Sebastian Moeller wrote: > Hi Aaron, > > > > On Mar 9, 2023, at 16:43, Aaron Wood via Cerowrt-devel < > cerowrt-devel@lists.bufferbloat.net> wrote: > > > > At least for the sub-1GHz 802.15

Re: [Cerowrt-devel] the network is devolving

10Mbit single pair would be for an entirely different application than the printer, though. That’s a single-pair power-and-data solution for a microcontroller/sensor network. 1km is good for that speed off of a single twisted pair. In my past experience that would have been 40-100kbps at most (i

<    1   2