This article references a WDS limitation and suggest MPLS as a more efficient way to brodge traffic across N wireless link:
http://wiki.mikrotik.com/wiki/Transparently_Bridge_two_Networks_using_MPLS
The WDS limitation has been mentioned in several other threads as well discussing wireless N throughput.
What exactly is the WDS limitation? Whats the max bandwidth you can squeeze through WDS link with two RB433AH boards each with R52N card setup however you want?
Thanks,
Scott
Due to how 802.11n technology works (some of it’s features), it doesn’t cooperate with WDS technology so well, this causes the throughput to be lower, this is why you can use other ways described in the Wiki
I understand this, what I am wondering is how much bandwidth I can get using WDS and N wireless…
10MB, 50MB, 100MB etc…
Once I know the limitation I can decide whether I need to go fster or can just use WDS the way it is with N and still have the BW I require.
Thanks,
Scott
ping
I know this is a couple of months late, but here is my feed back. With WDS I have seen a 50% decrease in throughput (realworld) from what the radio is connected at. I have a 4 mile point to point using ubiquiti and the radios are hooked up at -61 and have a speed of 117 Mbps both ways. Best speedtest to an on network speedtest (on gig backbone) is 45 Mbps.
On the other hand I did setup an mpls link like the instructions say and it is hooked up at 150 Mbps both ways. I can get about 100 Mbps one way on radio speed tests. When I do a real world test, the best I can get is 20 Megs.
Very strange.
-Z
Hi all,
Normis, why mikrotik wds on 802.11n has low bandwidth, but with ubnt airmax devices the wds on N works well, with no significant difference? I now that you cannot answer by ubnt, but their software engineering team “who are them???” managed a way for it to work well.
I seriously think that in the next months, mikrotik will have all this problems of Nstream and wds solved, but its taking to much time, and it starts to be difficult to arrange excuses to our customers.
Kindly regards,
Jorge
I got wireless n to work with wds. use wds-station mode and ap-bridge.
what speeds do you get with ubnt and with mikrotik? what are the differences in speed?
It seems like that WDS technology doesn’t work very well with MikroTik, not with 802.11n. 
If your using WDS as ptp it works great. I have done it with nstreme and 5.8. I dont have 802.11n yet.
Is this ever going to be fixed? I can only get 30 Mbps of advertised 300 Mbps.
WDS over every hop cuts the bandwidth in half. Could you explain your problem.
I have WDS-mesh between 2 hops (SXT 2nD r2, 2011UAS-2HnD), WPA2, distance 20 meters.
hop A:
0 interface=wlan1 radio-name="D4CA6DAB65F0" mac-address=D4:CA:6D:AB:65:F0 ap=yes wds=yes bridge=no rx-rate="300Mbps-40MHz/2S/SGI" tx-rate="270Mbps-40MHz/2S" packets=184208,156244
bytes=40267563,103560628 frames=184208,156246 frame-bytes=40637463,102623424 hw-frames=202139,156494 hw-frame-bytes=49075051,108887909 tx-frames-timed-out=0 uptime=19h31m11s
last-activity=10ms signal-strength=-46dBm@1Mbps signal-to-noise=56dB signal-strength-ch0=-50dBm signal-strength-ch1=-48dBm tx-signal-strength-ch0=-39dBm
tx-signal-strength-ch1=-33dBm strength-at-rates=-46dBm@1Mbps 40ms,-44dBm@2Mbps 19h31m11s100ms,-46dBm@5.5Mbps 19h31m10s970ms,-46dBm@11Mbps 19h31m10s900ms,-41dBm@6Mbps 19h31m11s190ms,-
40dBm@9Mbps 19h31m10s970ms,-42dBm@12Mbps 19h31m10s750ms,-40dBm@18Mbps 19h31m10s570ms,-40dBm@24Mbps 19h31m10s330ms,-43dBm@36Mbps 19h31m10s90ms,-43dBm@48Mbps
19h31m9s730ms,-44dBm@54Mbps 19h31m9s350ms,-42dBm@HT20-0 19h31m11s180ms,-41dBm@HT20-1 19h31m10s950ms,-41dBm@HT20-2 19h31m10s800ms,-43dBm@HT20-3 19h31m10s620ms,-
41dBm@HT20-4 19h31m10s440ms,-43dBm@HT20-5 19h31m10s200ms,-44dBm@HT20-6 19h31m9s720ms,-46dBm@HT20-7 19h31m9s420ms,-46dBm@HT40-0 19h31m11s170ms,-46dBm@HT40-1
19h31m10s950ms,-46dBm@HT40-2 19h31m10s780ms,-46dBm@HT40-3 19h31m10s600ms,-50dBm@HT40-4 7h57m16s590ms,-49dBm@HT40-5 2h59m17s310ms,-49dBm@HT40-6 1s170ms,-50dBm@HT40-7
20ms
tx-signal-strength=-32dBm tx-ccq=93% rx-ccq=100% p-throughput=251276 nstreme=no framing-mode=none routeros-version="6.34.3" 802.1x-port-enabled=yes authentication-type=wpa2-psk
encryption=aes-ccm group-encryption=aes-ccm management-protection=no compression=no wmm-enabled=yes tx-rate-set="CCK:1-11 OFDM:6-54 BW:1x-2x SGI:1x-2x HT:0-15"
hop B:
0 interface=wlan1 radio-name="D4CA6D973C30" mac-address=D4:CA:6D:97:3C:30 ap=yes wds=yes bridge=no rx-rate="300Mbps-40MHz/2S/SGI" tx-rate="300Mbps-40MHz/2S/SGI" packets=156695,184175
bytes=103599824,40362922 frames=156695,184177 frame-bytes=103914830,39258108 hw-frames=172732,184439 hw-frame-bytes=125522894,46638707 tx-frames-timed-out=0 uptime=19h32m36s
last-activity=280ms signal-strength=-29dBm@1Mbps signal-to-noise=81dB signal-strength-ch0=-37dBm signal-strength-ch1=-30dBm tx-signal-strength-ch0=-50dBm
tx-signal-strength-ch1=-47dBm strength-at-rates=-29dBm@1Mbps 50ms,-31dBm@2Mbps 19h32m35s500ms,-29dBm@5.5Mbps 19h32m35s400ms,-32dBm@11Mbps 19h32m35s300ms,-31dBm@6Mbps 19h32m35s580ms,-
27dBm@9Mbps 19h32m35s400ms,-32dBm@12Mbps 19h32m35s150ms,-32dBm@18Mbps 19h32m34s980ms,-30dBm@24Mbps 19h32m34s740ms,-34dBm@36Mbps 19h32m34s500ms,-34dBm@48Mbps
19h32m34s190ms,-35dBm@54Mbps 19h32m33s810ms,-29dBm@HT20-0 19h32m35s580ms,-27dBm@HT20-1 19h32m35s380ms,-25dBm@HT20-2 19h32m35s200ms,-29dBm@HT20-3 19h32m35s20ms,-
27dBm@HT20-4 19h32m34s850ms,-25dBm@HT20-5 19h32m34s610ms,-36dBm@HT20-6 1m34s620ms,-39dBm@HT20-7 1m57s70ms,-33dBm@HT40-0 19h32m35s570ms,-36dBm@HT40-1 19h32m35s380ms,-
29dBm@HT40-2 19h32m35s190ms,-31dBm@HT40-3 19h32m35s20ms,-31dBm@HT40-4 8h7s710ms,-38dBm@HT40-5 1m15s790ms,-39dBm@HT40-6 1m15s900ms,-42dBm@HT40-7 280ms
tx-signal-strength=-46dBm tx-ccq=95% rx-ccq=33% p-throughput=268109 nstreme=no framing-mode=none routeros-version="6.34.3" 802.1x-port-enabled=yes authentication-type=wpa2-psk
encryption=aes-ccm group-encryption=aes-ccm management-protection=no compression=no wmm-enabled=yes tx-rate-set="CCK:1-11 OFDM:6-54 BW:1x-2x SGI:1x-2x HT:0-15"
I do BTest from hop A to hop B:
status: done testing
duration: 1m1s
rx-current: 30.3Mbps
rx-10-second-average: 28.7Mbps
rx-total-average: 28.7Mbps
lost-packets: 367
random-data: no
direction: receive
rx-size: 1500
CPU load during test is below 60%. I expect rx-total-average near rx-rate/tx-rate of the link and lost-packets near zero.
The connection is clear and stable:
sent=100 received=100 packet-loss=0% min-rtt=1ms avg-rtt=1ms max-rtt=34ms
try udp test
Mmm… This was udp test (of course)/
spectrum analyze each ap. But keep in mind Nstreme is a lot faster. 802.11n is not that effective algorithm.
also learn to lock data rates.
Can I use Nstreme for Point-to-Point and 802.11 for regular client connection simultaneously?
They only way you can do that is if you do wds 5ghz then 2.4ghz to the client. For your set the answer is no.
Thanks! One more question: can AP work at different speeds simultaneously? F.ex. at 54m with client and at 300m P-to-P. Or it is the same restriction: one radio module – one tx speed?