ured IP rate) AND use the link layer adaption settings? (85-90% of
bandwidth)
or
- rely on the LLA settings to do the overhead, and shape to just a tiny bit
under the reported line rate? (95-99% of bandwidth)
Thanks,
- Aaron Wood
On Sat, Jan 11, 2014 at 8:06 PM, Rich Brown wrote:
> Hi Sebas
All,
I'm noting this here in case anyone is interested. After I write this up,
I'm going to start from scratch on the configuration, and factory-reset the
router.
=
The 5GHz radio on my 3800 seems to be in a very odd state. I'm not quire
sure what state it's in, but it seems to be only doi
/:00:12.0'
option htmode 'HT40+'
list ht_capab 'SHORT-GI-40'
list ht_capab 'TX-STBC'
list ht_capab 'RX-STBC1'
list ht_capab 'DSSS_CCK-40'
option txpower '17'
option country 'FR'
-Aaron
On Thu, Jan 16, 2014 at 4:
>
> > The 5GHz radio on my 3800 seems to be in a very odd state. I'm not
> quire sure what state it's in, but it seems to be only doing HT20 1x1. And
> in a fairly broken manner at that.
> >
> > Running the rrul test (over wifi directly to the router as the
> netserver), tcp uploads were 25Mbps o
>> Sebastian, after sorting out the router, it's still biased, but far
>> less
>> so, about a 2:1 ratio between upload and download.
>
> So I See offen 10:1 and worse @165Mbit/s raw wireless rate
I get mixed results, but they aren't good. IIRC, apple really changed
something about the media ac
On Thu, Jan 16, 2014 at 9:10 PM, Sebastian Moeller wrote:
> Hi Aaron,
>
> I just checked again and I get crazy results for both RRUL and
> RRUL_NOCLASSIFICATION:
>
Yes, those look a like my results (after having gotten things running).
When broken, it was still imbalanced, but the overall spe
It would be interesting to see what streamboost gains on top of just
fq_codel. Since it appears that they are doing so fairly heavy-handed
traffic shaping.
-Aaron
On Fri, Jan 24, 2014 at 12:48 PM, Dave Taht wrote:
> An interesting new chipset:
>
>
> http://www.anandtech.com/show/7526/qualcomm
>
> of the other commonly installed ones, there are samba, openwrt,
> strongswan and ipsec, that I know of. All these are packages that
> haven't been installed or tested for a while (by me, anyway)...
>
I have strongswan running on my 3800 with v3.10.24-8. I've found some
interesting results com
Thanks for sharing these, especially because it becomes clear how the various
traffic classes should behave (BK and EF in particular)
-Aaron
Sent from my iPhone
On Jan 25, 2014, at 21:52, Dave Taht wrote:
> Notes and graphs here:
>
> http://snapon.lab.bufferbloat.net/~cero2/jimreisert/result
On Mon, Feb 10, 2014 at 7:39 PM, Sebastian Moeller wrote:
> Hi Dave,
>
> On Feb 10, 2014, at 18:48 , Dave Taht wrote:
> > I don't think this has anything to do with the the ht40+ MCS15 problem
> at 5ghz.
>
I'm getting 40MHz (MCS13 at the moment, often MCS15) on ch 36, country-code
FR (but that
>
> I'm getting 40MHz (MCS13 at the moment, often MCS15) on ch 36,
> country-code FR (but that was expected, right?)
>
I'm also able to get 40MHz operation ch 44.
Putting the radio onto channel 40, however, disabled it:
Tue Feb 11 12:22:11 2014 daemon.notice netifd: radio1 (1046): Configuration
On Tue, Feb 11, 2014 at 2:34 PM, Aaron Wood wrote:
> Remembering that HT40+ and ch 48 weren't allowed, I tried to switch to
>> HT40-. And... now it's really borked up. The luci ui won't give me any
>> options other than client mode on 48 or auto. I'm go
Let me know when this is in a build, and I'll test it here.
-Aaron
On Sat, Feb 22, 2014 at 12:55 AM, Sebastian Moeller wrote:
> Hi Dave,
>
> so I reshuffled my todo list a bit and finished the exposure of limit and
> target in the GUI. For target, the string auto will use the curve you
> recen
Do you have the latest (head) version of netperf and netperf-wrapper? some
changes were made to both that give better UDP results.
-Aaron
On Mon, Feb 24, 2014 at 3:36 PM, Rich Brown wrote:
>
> CeroWrt 3.10.28-14 is doing a good job of keeping latency low. But... it
> has two other effects:
>
Rich,
One thing that helped considerably on my my DSL link (21000/1200), was
turning on the Link Layer Adaptation. With that, and efq_codel, I've been
very happy with the (nearly non-existent) latency. Yeah, I lost a couple
percent off the top, but the behavior is better. Although I was startin
Is there a writeup on each of the fq_codel variants?
-Aaron
Sent from my iPhone
> On Mar 2, 2014, at 22:41, Dave Taht wrote:
>
> Nice work!
>
> I have a problem in that I can't remember if target autotuning made it
> into that release or not.
>
> Coulde you do a tc -s qdisc show dev ge00 on
yes. I've been meaning to write this up. here's the quick version (as
it's very late in Paris):
I followed these instructions, but ran into issues:
http://www.tschofenig.priv.at/wp/?p=963
- fping3 is needed (it's in Homebrew)
- netperf doesn't compile by default with the clang compiler, the sta
>
> > The ntp servers queried presently largely are not dnssec signed, so
> > the ntp queries
> > should succeed (I think?) in the general case. However, for
> > robustness, I'd argue for enhancing the ntp startup script to
> > temporarily disable dnssec until it gets a valid time, and then
> > ena
On Sun, Mar 23, 2014 at 12:15 PM, Toke Høiland-Jørgensen wrote:
> Aaron Wood writes:
>
> > or we find a way to have long-lived dnssec entries.
>
> Is the timing controllable somehow? I.e. would it be possible to set up
> a special domain name with a really long-lived key
>
> The ., org. keys are not going to grow multiple year expiries, so we need
> our
> own thing to cache. One could cache the DNSKEY for bufferbloat.net along
> with the root zone keys... then lookup ntp.bufferbloat.net. It would have
> to
> return a A/ records, because chasing a CNAME into nt
Yesterday, I upgraded to 3.10.32-12 (rrul results coming soon, it's a big
improvement over 3.10.24-8, though I'm not sure why, exactly).
This morning I turned on dns request logging, since I wanted to keep a
better eye on the dnssec results (and so that I had some short-term
forensic trails in cas
Nice! It gives me what I'd expect for my setup, although the TCP rate is
only 2/3 the line rate on DSL (upload is better at 80%).
..
Download: 14.09 Mbps
Latency: (in msec, 62 pings, 0.00% packet loss)
Min: 30.157
10pct: 30
.tohojo.dk which is much closer to you.
>
> (ping times between there and rich's server would be good to have)
>
>
>
> On Tue, Mar 25, 2014 at 9:09 AM, Aaron Wood wrote:
> > Nice! It gives me what I'd expect for my setup, although the TCP rate is
>
I also don't consider the ntp/dnssec issue a blocker, not at the moment.
It's a larger problem to solve, and one that needs solving in a wider
context than just CeroWRT, and so we should keep working on a solution, but
not make it a "release blocking" issue. It's a known issue, a known bit of
res
Dave had asked about results from .32-12 on DSL, and in particular how pie
was fairing on dsl. I finally was able to setup a clean test env
yesterday, and ran a bunch of tests.
Results:
http://burntchrome.blogspot.com/2014/03/cerowrt-31032-12-sqm-comparison-on.html
Takeaways:
- I'm still droppi
On Wed, Mar 26, 2014 at 11:11 PM, David Lang wrote:
> If the openwrt folks could figure out how they are going to deal with NAND
> flash, it would be nice to be able to use one of the many routers that is
> shipping with more flash (128M in the newer netgear routers would be nice)
>
> if I were t
On Fri, Mar 28, 2014 at 9:36 AM, Mikael Abrahamsson wrote:
> On Thu, 27 Mar 2014, David Lang wrote:
>
> using a 3800 or similarly priced ($100-$150 USD) device that's readily
>> available is very good for the second category, the question is if we can
>> find one that's powerful enough for the fi
As a contra-datapoint:
4d uptime here, but without any ipv6. A couple GB of data, both
WAN<->internal, and between 5GHz devices (pushing to an AppleTV). No real
traffic between 2.4 and 5GHz, though.
WPA2 on all interfaces (babel and guest interfaces are disabled).
-Aaron
On Fri, Mar 28, 2014
On Fri, Mar 28, 2014 at 8:39 PM, Dave Taht wrote:
> On Fri, Mar 28, 2014 at 11:40 AM, Michael Richardson
> wrote:
> >
> > Mikael Abrahamsson wrote:
> > > I also think that it'll be very hard to find a box that is decently
> > > priced
> > > that also will do gig speeds *and* will do
On Tue, Apr 1, 2014 at 8:21 PM, Dave Taht wrote:
> On Thu, Mar 27, 2014 at 3:21 AM, Aaron Wood wrote:
> > Dave had asked about results from .32-12 on DSL, and in particular how
> pie
> > was fairing on dsl. I finally was able to setup a clean test env
> yesterday,
> &g
Up for 10 days on 3.10.32-12 (WNDR3800). Only have 2 devices that run
2.4GHz, and it's only seen 2GB of traffic on SW00 in that time... The 5GHz
radio has had >5GB of traffic on it in the same time. No problems at all.
-Aaron
On Fri, Apr 4, 2014 at 12:51 AM, Maxim Kharlamov wrote:
> The las
On Fri, Apr 4, 2014 at 12:56 AM, Aaron Wood wrote:
> Up for 10 days on 3.10.32-12 (WNDR3800). Only have 2 devices that run
> 2.4GHz, and it's only seen 2GB of traffic on SW00 in that time... The 5GHz
> radio has had >5GB of traffic on it in the same time. No problems at a
On Fri, Apr 4, 2014 at 12:58 AM, Dave Taht wrote:
> On Thu, Apr 3, 2014 at 3:57 PM, Aaron Wood wrote:
> > On Fri, Apr 4, 2014 at 12:56 AM, Aaron Wood wrote:
> >>
> >> Up for 10 days on 3.10.32-12 (WNDR3800). Only have 2 devices that run
> >> 2.4GHz, and i
David,
I've been using both nfq and fq with a 16/1 DSL line, and never run into
it, but most of my traffic is on the 5GHz radio, not the 2.4GHz radio.
-Aaron
On Fri, Apr 11, 2014 at 11:50 AM, David Personette wrote:
> I just thought of it now, but at Dave's suggestion (because of having a
> DSL
>
> So far as I know the caching functionality in dnsmasq in that instance
> is disabled due to fears about cache poisoning, that I don't fully
> understand. My half understood fear translates into equivalent fears
> for other local dns daemons.
Which isn't near the issue that application-level c
cero 3.10.32-12, on a 3800
I had a power outage (power strip bumped), and the wndr rebooted, and came
up without wireless radios running. I ssh'd in and found this in the logs:
Wed Mar 26 14:33:41 2014 user.notice root: ingress shaping activated
Wed Mar 26 14:33:41 2014 kern.info kernel: [ 27.
On Tue, Apr 15, 2014 at 9:49 PM, Dave Taht wrote:
> There were several segfault bugs fixed in netifd since that release.
> There was also an issue with correctly parsing the wireless regdb.
> Please upgrade.
What's the recommended version?
> You've got heartbleed, too.
>
Only visible to me,
In about 2 months, I'll have my hands on some HomePlugAV devices that I can
do extensive testing with. My previous tests were showing very, very large
bufferbloat, when used on poor links (2000ms of buffering and growing over
a 1-2 minute rrul test that wasn't able to get over 10Mbps due to SF
Vic
One of the many servers involved with BofA's online banking:
Sat Apr 19 09:37:37 2014 daemon.info dnsmasq[29719]: using nameserver
8.8.4.4#53
Sat Apr 19 09:37:37 2014 daemon.info dnsmasq[29719]: using nameserver
8.8.8.8#53
Sat Apr 19 09:37:37 2014 daemon.info dnsmasq[29719]: using local addresses
I'm setting up new service in the US, and I'm currently assuming that all
of Comcast's rates are "boosted" rates, not the "provisioned" rates.
So if they quote 50/10Mbps, I assume that's not what will need to be set in
SQM with CeroWRT.
Does anyone have good info on the "provisioned" rates that g
s of the numbers?
> It's probably too much to ask for Comcast to go on the record with a
> precise definition.
>
>
>
>
>
> On Saturday, April 19, 2014 8:55am, "Aaron Wood" said:
>
> I'm setting up new service in the US, and I'm currently assuming that
pear that this is
>> indeed a bogus DS.
>>
>> http://dnssec-debugger.verisignlabs.com/sso-fi.bankofamerica.com
>>
>>> On Sat, Apr 19, 2014 at 2:43 AM, Aaron Wood wrote:
>>> One of the many servers involved with BofA's online banking:
>>>
&
Wed Apr 23 15:13:05 2014 daemon.info dnsmasq[29719]: query[A]
e3191.dscc.akamaiedge.net.0.1.cn.akamaiedge.net from 172.30.42.99
Wed Apr 23 15:13:05 2014 daemon.info dnsmasq[29719]: forwarded
e3191.dscc.akamaiedge.net.0.1.cn.akamaiedge.net to 8.8.8.8
Wed Apr 23 15:13:05 2014 daemon.info dnsmasq[2971
On Wed, Apr 23, 2014 at 6:44 PM, Robert Bradley
wrote:
>
> > ; <<>> DiG 9.8.1-P1 <<>> +cd @8.8.8.8 a
> > e3191.dscc.akamaiedge.net.0.1.cn.akamaiedge.net
>
> >
> > But a query for DS on the same domain, which is what dnsmasq does next,
> > returns SERVFAIL, _even_with_ checking disabled.
> >
> > ;
On Wed, Apr 23, 2014 at 5:58 PM, Simon Kelley wrote:
> On 23/04/14 16:42, Dave Taht wrote:
> > I will argue that a better place to report dnssec validation
> > errors is the dnsmasq list.
> >
> > On Wed, Apr 23, 2014 at 8:31 AM, Aaron Wood wrote:
> >>
Well, I'm seeing the same results as you are from here in Paris (using
Free.fr).
-Aaron
On Thu, Apr 24, 2014 at 1:27 PM, Simon Kelley wrote:
> On 24/04/14 11:49, Aaron Wood wrote:
>
> >
> >> Dnsmasq does the DS query next because the answer to the A query comes
>
Using CeroWRT 3.10.36-4, I'm seeing the following in the logs:
Thu Apr 24 14:15:14 2014 daemon.info dnsmasq[13365]: query[PTR]
b._dns-sd._udp.96.42.30.172.in-addr.arpa from 172.30.42.99
Thu Apr 24 14:15:14 2014 daemon.info dnsmasq[13365]: forwarded
b._dns-sd._udp.96.42.30.172.in-addr.arpa to 8.8.8
Just too many sites aren't working correctly with dnsmasq and using
Google's DNS servers.
- Bank of America (sso-fi.bankofamerica.com)
- Weather Underground (cdnjs.cloudflare.com)
- Akamai (e3191.dscc.akamaiedge.net.0.1.cn.akamaiedge.net)
And I'm not getting any traction with reporting the errors
ocess is going to be more tricky, because
> dnsmasq is easily dismissed as too small a proportion of the market to
> care. (wish it were not so).
>
>
>
> On Saturday, April 26, 2014 7:38am, "Aaron Wood" said:
>
> Just too many sites aren't working correctly
This timeout, I'm guessing this is older/naive setups that aren't expecting
to support DNSSEC, and thought "over-securing" their setup, have managed to
break the non-existence-proof process?
-Aaron
On Mon, Apr 28, 2014 at 9:32 PM, Simon Kelley wrote:
...
> Neither of authoritative nameservers f
I'm doing some tests (playing with some new tests for netperf-wrapper), and
my usual test server appears to be down (Toke's server). And Dave's server
at bufferbloat.net is _way_ too far away at 180ms (and I get really uneven
bandwidth to there).
And feel free to reply off-list if you don't want
On Tue, Apr 29, 2014 at 5:46 PM, Dave Taht wrote:
> > Yes, but as soon as you hit the long distance network the latency is the
> > same regardless of access method. So while I agree that understanding the
> > effect of latency is important, it's no longer a meaningful way of
> selling
> > fiber
Now that I'm on Comcast, I'm going to try it again.
-Aaron
On Fri, May 16, 2014 at 8:25 PM, Stephen Hemminger <
step...@networkplumber.org> wrote:
> On Sat, 26 Apr 2014 13:38:08 +0200
> Aaron Wood wrote:
>
> > Just too many sites aren't working correctly
I _think_ it was the comcast keyword. That or a +1 by Yonatan Zunger
(Google+ lead), which got implicitly shared to a wider audience... I'm
glad to have gotten us some visibility into this (I started revising for
clarity as the exposure increased).
btw, I'm about to do those tests with PIE you a
http://burntchrome.blogspot.com/2014/05/i-dont-like-this-pie-more-bufferbloat.html
PIE certain is better than nothing. But it's no fq_codel.
All I varied between the two is the queuing discipline, with the PIE target
(both directions) set to 16ms.
not impressed.
But, far better than pie did wh
It depends on the aqm rules that are configured. In the base setup, it
struggles at 50Mbps. But that can be increased by switching from the
simple.qos script to simplest.qos (I'm not sure where the limit is with the
simplest.qos script.
I know that Dave Taht has been working with some other plat
http://www.gateworks.com/product/item/ventana-gw5310-network-processor
Out of price range in single units, but I don't know where the price breaks
kick in. Dual-core 800MHz ARM should be plenty of power for the GigE ports.
I think to get any sort of platform like this by a major vendor, we're
lo
Comcast has upped the download rates in my area, from 50Mbps to 100Mbps.
This morning I tried to find the limit of the WNDR3800. And I found it.
50Mbps is still well within capabilities, 100Mbps isn't.
And as I've seen Dave say previously, it's right around 80Mbps total
(download + upload).
ht
On Fri, Aug 29, 2014 at 11:06 AM, Dave Taht wrote:
> On Fri, Aug 29, 2014 at 9:57 AM, Aaron Wood wrote:
> > Comcast has upped the download rates in my area, from 50Mbps to 100Mbps.
> > This morning I tried to find the limit of the WNDR3800. And I found it.
> > 50Mbps
> > But this doesn't really answer the question of why the WNDR has so much
> lower a ceiling with shaping than without. The G4 is powerful enough that
> the overhead of shaping simply disappears next to the overhead of shoving
> data around. Even when I turn up the shaping knob to a value quite
Luckily, I don't mind being wrong (or even _way_ off the mark).
I don't think that's it.
>
> First a nitpick: the PowerBook version of the late-model G4 (7447A)
> doesn't have the external L3 cache interface, so it only has the 256KB or
> 512KB internal L2 cache (I forget which). The desktop ver
>
> What this makes me realize is that I should go instrument the cpu stats
> with each of the various operating modes:
>
> * no shaping, anywhere
> * egress shaping
> * egress and ingress shaping at various limited levels:
> * 10Mbps
> * 20Mbps
> * 50Mbps
> * 100Mbps
>
So I set th
On Wed, Sep 3, 2014 at 4:08 AM, Jonathan Morton
wrote:
> Given that the CPU load is confirmed as high, the pcap probably isn't as
> useful. The rest would be interesting to look at.
>
> Are you able to test with smaller packet sizes? That might help to
> isolate packet-throughput (ie. connecti
I'm up to 12 days at this point, but it was longer (a month?) before I did
some re-arranging of things and needed to unplug the unit.
I have an AppleTV that's wired, and all the computers are on wifi, so we
end up stressing the mdns a fair bit. It seems solid, although all the
devices get confuse
On Mon, Oct 20, 2014 at 3:48 PM, Dave Taht wrote:
> OK, I tried a few combinations of burst and cburst on a cerowrt box,
> using 90/10 as a up/download speed.
>
> burst 64000 cburst 64000 was a bit of a win, in most respects, but odd
> in others.
>
Interesting can this be specified on the a
Uptime: 21:30:42 up 40 days, 12:17, load average: 0.11, 0.13, 0.08
no traps
It's running native IPv6 (Comcast), now that I have the Comcast Wifi unit
running in bridge mode. Everything seems just fine.
df -h:
FilesystemSize Used Available Use% Mounted on
rootfs
I have a couple friends in that crowd, and they _also_ aren't using shared
lines. So they don't worry in the slightest about congestion when they're
trying to keep dedicated links fully saturated. They're big issue with
dropped packets is that some of the TCP congestion-control algorithms kick
in
It's why I use Blogger for writeups of the netperf testing, and then post
to G+. The lack of inline images in G+ is just too limiting for that sort
of write-up.
-Aaron
On Sun, Nov 23, 2014 at 9:16 AM, Dave Taht wrote:
> I setup a bunch of picostations running openwrt barrier breaker to try
> a
Do I win any prizes? :)
root@cerowrt:~# uptime
16:27:49 up 74 days, 7:14, load average: 0.46, 0.17, 0.09
root@cerowrt:~# free total used free
shared buffers
Mem:126256496527660400
-/+ buffers: 49652
So it appears that Marvell pushed a bunch of patches to OpenWRT on
Christmas, and as a result, trunk OpenWRT (kernel 3.18) can run, and run
pretty well, on the 1900AC.
https://forum.openwrt.org/viewtopic.php?pid=258634#p258634
dual-core 1.2GHz ARM. Given my experience with processors like the iM
On Tue, Jan 6, 2015 at 11:37 AM, Dave Taht wrote:
>
> I also tend to wish that streaming video had got it's own control port
> rather than being layered over 80 and 443.
>
In my experience, that was due to the corporate firewalls' default rule of
disallowing outbound connections. Port 80 can be
On Sun, Jan 25, 2015 at 12:17 PM, wrote:
> There is no reason why one cannot set up an enterprise network to support
> roaming, yet maintaining the property that IP addresses don't change while
> roaming from AP to AP. Here's a simple concept, that amounts to moving
> what would be in the Ethern
On Fri, Feb 6, 2015 at 12:47 PM, Dave Taht wrote:
> On Sat, Feb 7, 2015 at 8:42 AM, wrote:
> > On Fri, 06 Feb 2015 15:27:32 +1300, Dave Taht said:
> >> so, how's everybody's uptime?
> >
> > Sitting at 27 days due to a power blip.
>
> I do strongly feel that home routers should have a battery or
I would love to see this data, especially since insights into dense wifi
environments is rare.
It seems like this might be a fascinating time to get stats from Minstrel,
for which rates clients are connected at:
Do they connect at high rates at all?
Do they stay stable at high rates, or do they s
Perhaps just a wall of shame? No venom, just point out the failings, and
call people out.
But, frankly, I don't think any of the router mfr's actually care (I've
seen no evidence of it), and since they're not in the business of actually
making these things (just putting their labels on them), I d
But until the silicon vendors update _their_ forks of OpenWRT, commercially
available home routers won't have these benefits. Because the home router
market is dominated by packaged reference designs from one of a very small
number of companies that actually make all the chipsets (Dave, I know you
I would definitely be interested in being involved with how to secure and
firewall, but still provide access to, internal IPv6 hosts. Ie, the
internet the way it's supposed to work (peer to peer), but with the
security that we've inadvertently picked up along the way by using NAT
everywhere for th
Bill,
I'd recommend setting the bandwidth values low (very low) at first, just to
establish that the setup is working correctly. I'm able to get better
control of latency at those bitrates on an WNDR3800:
http://burntchrome.blogspot.com/2014/05/fixing-bufferbloat-on-comcasts-blast.html
I'd star
On Thu, Mar 19, 2015 at 9:01 AM, Richard Smith wrote:
> On 03/16/2015 01:49 PM, David Lang wrote:
>
>> On Mon, 16 Mar 2015, Dave Taht wrote:
>>
>> On Mon, Mar 16, 2015 at 10:34 AM, David Lang wrote:
>>>
>>> On Mon, 16 Mar 2015, Dave Taht wrote:
1) We need a new box that can do inbou
I do this often at work, using a separate machine to capture traffic using
wireshark. Wireshark makes a lot of the analysis fairly straightforward
(especially with it's excellent packet dissectors). By capturing in
radiotap mode, you get RSSI/noise levels on the 802.11n packet, the rates
involved
On Fri, Mar 27, 2015 at 8:08 AM, Richard Smith wrote:
> Using horst I've discovered that the major reason our WiFi network sucks
> is because 90% of the packets are sent at the 6mbit rate. Most of the rest
> show up in the 12 and 24mbit zone with a tiny fraction of them using the
> higher MCS ra
OpenWRT normally sets it in the Build Options section of the menuconfig.
On Tue, Mar 31, 2015 at 10:20 AM, Dave Taht wrote:
> How do I try a different set of compile options nowadays?
>
> I change it in CC .config, it changes back.
>
> Used -O2 in cero...
>
> python, for example, is way slower t
code unit-tests with mocked clocks catch a lot of funky things like this.
But you have to either know to test a piece of code, or be disciplined
about testing _everything_ like that. And it doesn't retrofit into
existing codebases well at all.
On Sat, May 2, 2015 at 9:42 AM, Dave Taht wrote:
>
ICMP prioritization over TCP?
> >Ideas?
>
___
Cerowrt-devel mailing list
Cerowrt-devel@lists.bufferbloat.net
https://lists.bufferbloat.net/listinfo/cerowrt-devel
All,
I've been lurking on the OpenWRT forum, looking to see when the CC builds
for the WRT1900AC stabilized, and they seem to be so (for a very "beta"-ish
version of stable).
So I went ahead and loaded up the daily ( CHAOS CALMER (Bleeding Edge,
r45715)).
After getting Luci and sqm-scripts insta
May 23, 2015 at 10:17 PM, Aaron Wood wrote:
> All,
>
> I've been lurking on the OpenWRT forum, looking to see when the CC builds
> for the WRT1900AC stabilized, and they seem to be so (for a very "beta"-ish
> version of stable).
>
> So I went ahead and loaded u
On Sat, May 23, 2015 at 11:44 PM, Dave Taht wrote:
>
> And it has a fan. Hate fans. Amusingly (I guess), I had this same
> chipset to fiddle with in the "mirabox" and it ran wy too hot.
>
I haven't hit the fan, yet
> It is not clear why you are getting an inaccurate rate out of it, eit
I wrote this up on my blog, where I can intersperse text and graphs a bit
better:
http://burntchrome.blogspot.com/2015/06/htb-rate-limiting-not-quite-lining-up.html
Basically, I ran a series of tcp_download tests, using increasing ingress
rates with sqm_scripts, and then used flent's box-plots to
> > On the 3800, it never meets the rate, but it's only off by maybe 5%.
>
> As Jonathan pointed out already this is in the range of the
> difference between raw rates and tcp good put, so nothing to write home
> about ;)
>
Yeah, I'm not too worried about that 5%, based on that explanation
On Wed, Jun 3, 2015 at 3:27 PM, Dave Taht wrote:
>
>
>> kbps = quantum = time
>> 2 = 3000 = 1.2ms
>> 3 = 6000 = 1.6ms
>> 4 = 12000 = 2.4ms
>> 5 = 24000 = 3.84ms
>> 6 = 48000 = 6.4ms
>> 8 = 96000 = 9.6ms
>>
>
>
>> So it appears that the goal of these values was to keep incr
The buried-under-NDA datasheets _sometimes_ go into that detail, but it's
still usually pretty obtuse reading. And of course, it's all under NDA,
not part of the public datasheet.
-Aaron
On Mon, Jun 8, 2015 at 9:52 AM, Dave Taht wrote:
> have long hoped there was a way to do hardware egress ra
Here's a long-range (indoors) test result with the 1900AC. I have about
9dB snr at the client (all the way across my house, which has plaster
walls):
http://www.dslreports.com/speedtest/737736
download buffering is the 1900ac, upload buffering is the mac laptop's
driver. Both are trying too har
On Wed, Jun 24, 2015 at 8:07 PM, Mikael Abrahamsson
wrote:
> On Wed, 24 Jun 2015, Aaron Wood wrote:
>
> Here's a long-range (indoors) test result with the 1900AC. I have about
>> 9dB snr at the client (all the way across my house, which has plaster
>> walls):
&
On Fri, Oct 23, 2015 at 9:10 AM, Richard Smith wrote:
> I have a shiny new Linksys WRT1900ACS to test.
>
> I thought it might be nice to start with some comparisons of factory
> firmware vs OpenWRT with sqm enabled.
>
Here are my results with the same router (unless the WRT1900ACS is
different f
On Fri, Oct 23, 2015 at 1:18 PM, Richard Smith wrote:
> On 10/23/2015 02:41 PM, Michael Richardson wrote:
>
>> Richard Smith wrote:
>> > My test setup:
>>
>> > Laptop<--1000BaseT-->DUT<--1000baseT-->Server
>>
>> So, given that the DUT is the only real constraint in the network, what
>>
I've done the same setup in the past with my 3800, and htb limits just fine
to 10Mbps even when used with gigabit lab links.
So I think that, for whatever reason, htb just isn't functioning.
Dumping the qdiscs setup and stats using tc should make it clearer as to
what the state of things actually
Consul is based on Raft, so anyone using Consul is using Raft.
(and we're poking around at it at my company, but I don't have any insight
to give on it, yet). But in general, I also like distributed redundancy
(as opposed to primary/backup redundancy).
-Aaron
On Mon, Jan 25, 2016 at 9:40 AM, Da
I've often wanted the same thing: What's the time-length of given packets
(using various transmission rates), and the inter-packet delays, etc. What
_is_ 100% channel utilization, in terms of packets per second of a given
size/rate?
>From a pcap file full of radio-tap-level packets, can the chan
It will be interesting to see if that includes APs or just clients. The
OEMs are going to lose one of their big levers on price, if they lose
Broadcom from the mix (and Qualcomm is going to make some good money).
Although Marvell's chunk of that space has been growing...
-Aaron
On Fri, Mar 18, 2
Un-bloated power-line-to-AP units would be awesome. As would power-line to
POE adapters for small electronics. Although you have the same difficulty
with on-boarding there that you do with wifi.
-Aaron Wood
On Mon, Apr 18, 2016 at 9:35 AM, Jonathan Morton
wrote:
>
> > On 18 Apr, 20
1 - 100 of 138 matches
Mail list logo
