stablizing latency over wireless

hi guys
I’m running on nstreme dual on 6Km < range with 2 11g card, on 2.8.18
however my ping result is not constant and it can be much better.
any idea about the ack time or any configuration that I must setup.

I’m sure it is possible to achieve a constant response time over this link because trying other AP and bridge could get 1ms at all time however the actual bandwidth was much lower using other products…


Thanks

here is the ping result:

Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

C:\Documents and Settings\Administrator>ping 192.168.1.26 -t

Pinging 192.168.1.26 with 32 bytes of data:

Reply from 192.168.1.26: bytes=32 time=9ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time=2ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time=82ms TTL=64
Reply from 192.168.1.26: bytes=32 time=31ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time=39ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time=84ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time=17ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time=98ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time<1ms TTL=64
Reply from 192.168.1.26: bytes=32 time=1ms TTL=64

Ping statistics for 192.168.1.26:
    Packets: Sent = 23, Received = 23, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 0ms, Maximum = 98ms, Average = 15ms
Control-C
^C

here is my interface export:

/ interface ethernet 
set ether1 name="ether1" mtu=1500 arp=enabled disable-running-check=yes auto-negotiation=yes \
    full-duplex=yes long-cable=no speed=100Mbps disabled=no 
/ interface wireless 
set WLan24 name="WLan24" mtu=1500 arp=enabled disable-running-check=no radio-name="Athenaeum" \
    mode=nstreme-dual-slave ssid="PersiaSYS" frequency=2462 band=2.4ghz-b/g scan-list=default-ism \
    rate-set=default supported-rates-b=1Mbps,2Mbps,5.5Mbps,11Mbps \
    supported-rates-a/g=6Mbps,9Mbps,12Mbps,18Mbps,24Mbps,36Mbps,48Mbps,54Mbps basic-rates-b=1Mbps \
    basic-rates-a/g=6Mbps max-station-count=2007 ack-timeout=dynamic tx-power=default \
    noise-floor-threshold=default periodic-calibration=default burst-time=disabled fast-frames=yes \
    dfs-mode=none antenna-mode=ant-a wds-mode=static wds-default-bridge=bridge1 wds-ignore-ssid=yes \
    update-stats-interval=disabled default-authentication=yes default-forwarding=yes hide-ssid=no \
    802.1x-mode=none disabled=no 
set WLan58 name="WLan58" mtu=1500 arp=enabled disable-running-check=no radio-name="00095BE7209B" \
    mode=nstreme-dual-slave ssid="PersiaSYS" frequency=2437 band=2.4ghz-g-turbo \
    scan-list=default-ism rate-set=default supported-rates-b=1Mbps,2Mbps,5.5Mbps,11Mbps \
    supported-rates-a/g=6Mbps,9Mbps,12Mbps,18Mbps,24Mbps,36Mbps,48Mbps,54Mbps basic-rates-b=1Mbps \
    basic-rates-a/g=6Mbps max-station-count=2007 ack-timeout=dynamic tx-power=default \
    noise-floor-threshold=default periodic-calibration=default burst-time=1600 fast-frames=yes \
    dfs-mode=none antenna-mode=ant-a wds-mode=static wds-default-bridge=bridge1 wds-ignore-ssid=yes \
    update-stats-interval=disabled default-authentication=yes default-forwarding=yes hide-ssid=no \
    802.1x-mode=none disabled=no 
/ interface wireless nstreme 
set WLan24 enable-nstreme=yes enable-polling=yes framer-policy=dynamic-size framer-limit=3200 
set WLan58 enable-nstreme=no enable-polling=no framer-policy=best-fit framer-limit=3200 
/ interface wireless security 
set WLan24 security=none algo-0=none key-0="" algo-1=none key-1="" algo-2=none key-2="" algo-3=none \
    key-3="" transmit-key=key-0 sta-private-algo=none sta-private-key="" \
    radius-mac-authentication=no 
set WLan58 security=none algo-0=none key-0="" algo-1=none key-1="" algo-2=none key-2="" algo-3=none \
    key-3="" transmit-key=key-0 sta-private-algo=none sta-private-key="" \
    radius-mac-authentication=no 
/ interface wireless nstreme-dual 
add name="nstreme1" mtu=1500 arp=enabled disable-running-check=no tx-radio=WLan24 rx-radio=WLan58 \
    remote-mac=00:20:ED:0F:F7:00 tx-band=2.4ghz-g tx-frequency=2462 rx-band=2.4ghz-g-turbo \
    rx-frequency=2437 rates-b=1Mbps,2Mbps,5.5Mbps,11Mbps \
    rates-a/g=6Mbps,9Mbps,12Mbps,18Mbps,24Mbps,36Mbps,48Mbps,54Mbps framer-policy=exact-size \
    framer-limit=4000 disabled=no 
/ interface wireless align 
set frame-size=300 active-mode=yes receive-all=no audio-monitor=00:00:00:00:00:00 \
    filter-mac=00:00:00:00:00:00 ssid-all=no frames-per-second=25 audio-min=-100 audio-max=-20 
/ interface bridge 
add name="bridge1" mtu=1500 arp=enabled forward-protocols=ip,arp,appletalk,ipx,ipv6,other stp=yes \
    priority=32768 ageing-time=5m forward-delay=15s garbage-collection-interval=4s hello-time=2s \
    max-message-age=20s disabled=no 
/ interface bridge port 
set ether1 bridge=bridge1 priority=128 path-cost=10 
set WLan24 bridge=bridge1 priority=128 path-cost=10 
set WLan58 bridge=bridge1 priority=128 path-cost=10 
set nstreme1 bridge=bridge1 priority=128 path-cost=10 
/ interface l2tp-server server 
set enabled=no mtu=1460 mru=1460 authentication=mschap2,mschap1,chap,pap default-profile=default 
/ interface pptp-server server 
set enabled=no mtu=1460 mru=1460 authentication=mschap2,mschap1 keepalive-timeout=30 \
    default-profile=default 
[admin@PersiaSYS] interface>

i’ve found personally that with CM9 cards atleast winxp ping gives false reports, i just use linux ping and smokeping - gives a much better view of the link stats. Windows would report strange spikes while smokeping and linux give a stable and lower ping result.

example would be a nstream1 link doing 36mbit about 3km, windows reposrts ping of 1-2ms and spikes to 20, linux and smokeping report .7ms ping with spikes of 1.3ms - checking this in games and the such in windows confirms this.

YMMV

I’m not quite sure what exactly do you mean, but I tried pinging from three different platfrom (MT itself, Windows and Linux)

does this affect link speed or something else? as i see the difference (spikes) is very small.

Nail the radios to a fixed rate both ends?

I think as Stephan Patrick says: Lock the data rates of the card to te best one

normis, as a matter fact I thought that speed is okay which it is okay but only in Rx means when I download something it is almost 19~20 mb per second

but when I upload, it jumps to 20 mb for 1 or 2 seconds and then fall to 30kb !!! and then again and again…


thanks

here is the bandwidth graph:

actually I want do that but how to lock the Rx and Tx speed at different speed? since I am using Nstreme dual

thanks

You look under wireless interface config.
Have you considered that the card might be faulty?
Did you wear proper anti-static gear when you assembled the router?

no I didn’t wear anti static gear and I dont even know what is that,
but I had a faulty card before which I changed it but how do I check
which one of these 4 cards are faulty?

can you double check my configuration? is there anything wrong in
my settings?

thanks

2.8.18 is pretty old…have you tried 2.8.28? The latest 2.8 version has been stable for us as far as wireless goes. Now if the’d only fix…

By replacing them and see if it helps.
Perhaps you have a short in the antenna cables?

yes I have a short antenna cable, I tried to change some packet setting such as best fit and exact and the size and it stayed a bit longer at max speed, my guess is it can be also the transmission settings.

the Tx-Rx link that is working fine is actually on a converter from 2.4 to 5.8 so its channel is very free, but the Rx-Tx that has a problem is actually only on 2.4 with interference with at least 10 more AP :frowning:

isn’t it the cause ? the interference?

I meant a short IN the cable.
About the ten other AP’s on that same frequency (or close to), sure, that could very very easily do it. But why not use a 5.8ghz card instead? A converter could add noise but it also reduces rx/tx sensitivity/strength.
I bet that converter cost more than a 5.8ghz card!

The chances of the problem being interference is very high. There is a multiplicity of things that can cause interference on wireless equipment. I would recommend what was mentioned before, lock the devices to a set frequency and rate and see if that helps. Its always good to frequency hop to find the best one for you and lock it down. If this doesnt help, it sounds like a hardware issue.

Do you see the AP changing speeds? In my experience these ping spikes are typical of the AP constantly negotiating new speeds, even when the link is stable. This is in 802.11a mode, I have not use NStream.

By the way, I beleive that you cannot set a RX speed, only TX speed. So to set a RX speed you need to set the TX speed of the radio which is communicating with it.

AFAIK you can nail the radio rates at BOTH ends of a P2P, or at all CPEs and the AP in P2MP.
Otherwise in MT you quite often see the modulation rates show up as “54/6” or similar indicating that the uplink and downlink are at different rates.

Regards

it is true that my connection is like jumping from 54 to 48 (more at 54) mb in one way and 12~18 at the other way…

which I may as well think about the cable issue… here in malaysia is hard to find fine piece of connectors… and very expensive to ship from USA.

but the interesting thing the link with 12~18 and less signal streght as more stable transmission… and at the other hand when I lower the transmission to something like 2 ~ 4 mb it will be okay at upload transmit, and it will not flactuate from 20mb to 0 like the posted screen-shot… but when it passes 2~4 its the same again…

I tried to lock the speed but it was same ping time and same speed issue again… I tried to change the framepolicy and size and it did kept in high speed for longer time before falling to zero. what setting is the best for this issue?

Imagine the line with lower signal strenght but a noise free channel (5.8) is doing better with higher signal strenght and noisy channel. and for the frame policy and size there is no option to set separate values for Tx and Rx

I’m totally blank about how to fix this thing.

What version are you using, what band? Are using nstreme?
Have you contacted the support@mikrotik.com?
The rates usualy will stay at the lowest rate when there is no data going. You can try to do some bandwidth test through the link and look how high will the rates go. If it is constantly jumping from 54 to 48, then I suggest you to lock the rates from 6 till 48 to make the link stable.
It is normal that the rates on other way stays low - probably there is no data going through that direction.
Also if you are using nstreme in area which has lot of interference the link could be not so stable.

uldis - Why does the radio switch rate depending on the speed at which data is passing over the link? Why doesn’t it just negootiate the best rate based on error rate and then stay there until the error rate changes?

This would seem to be the root of the high jitter problems that I and many others are seeing.