As per previous discussion, we now have the bridge configured using LEAP/WPAv2. However, we are not seeing expected speeds when compared to another 'n' bridge we have using 2x 1142 units.
When transferring a file from a PC on one side of the 1532 bridge to another PC on the other side of the bridge we see it settle between 3 MB\s - 3.1 MB\s
root - 1532e with AIR-ANT2588P3M directional antenna (connected to ports 1 and 2 in single band mode, using 2x 5 meter n-type male extensions)
non root - 1532i
RSSI - consistent -39 to -40 dBm
When transferring a file from a PC on one side of the 1142 bridge to another PC on the other side of the bridge we see it settle between 7 MB\s - 8 MB\s
I've carried the tests out above numerous times and get the same results pretty much everytime. Speeds do spike at the start but the speeds above are the most consistent speeds I see.
Both bridges are currently using 2.4ghz interfaces to keep things simple.
Any advice would be great as these need to be up and running for the weekend :(
Thanks in advance
Do you have access to s spectrum analyzer to see if there are interference spikes? Or a wifi packet capture? These could be useful if you have them.
How are the devices connected? Are they direct from bridge to PC or is it through a networked switch? I have seen issues where bridges were placed on trunked ports without pruning and it caused issues with the broadcast traffic.
Your associated data rate is m15 or ~145mbps on a 20mhz channel with a single SS. The 1532e has 2 SS so you should see 300 with 20mhz and 600 with 40mhz. 5ghz is where you will want to be to jump to 40mhz.
I would see no reason not to use 40mhz channel bonding for your link unless you fear you may need the channels for other things.
Thanks for responding so quickly.
In the last hour I removed all switches from the equation so have done tests with this setup ..
PC (fixed IP) >> 1532i >> [bridge] >> 1532e >> PC (fixed IP)
Still the exact same speeds as previously and tops out around 3.1 MB\s consistently.
I've checked the status page on the 1532 vs status page on our 1142 bridge and they are pretty much identical (including the current rate showing @ m15).
I'm really hoping this isnt a bug in the new IOS on these 1532's :(
|Association: Station View- Client|
After obtaining a downlaod for the latest firmware from Cisco. ap1g3-k9w7-tar.153-3.JAB.tar. This resolved the throughput issue. Thanks for all help.
Hi, yes, upgrading the firmware resolved the issue. As posts above point out, Cisco found a bug that was causing the reduced throughput. We were able to acquire firmware from Cisco support directly; even though we purchased these 6 months prior via a third party, but we also had to reference this post i created - to confirm that the issue had occured since they were purchased.
Would you please email me @firstname.lastname@example.org which AP IOS code resolved it. I have the same issue and i need you to help me or share me the ios.
For a test move the bridge across to 5GHz and do the channel bonding to 40MHz like below.
channel width 40-above
If you run the below on the root bridge CLI, you will get the output to confirm the data rate bridge associated.
show dot11 associations <nonroot-bridge-mac-address>
Do the same file transfer test & compare the difference.
**** Pls rate all useful responses ****
We have made the change to 5GHz (over 40MHz), external antenna is connected to top ports 3 and 4 on the 1532e and is running in single band mode.
Transfer speeds are as follows after approx 30 mins of testing ...
PC (fixed IP) >> 1532i >> [bridge] >> 1532e >> PC (fixed IP) .... average 3.7 MB\s
PC (fixed IP) >> 1532e >> [bridge] >> 1532i >> PC (fixed IP) .... average 5.3 MB\s
Association stats from the root bridge (1532e) ..
Both configs for the root and non-root unit are also attached as text files.
Thanks in advance!
Also, to add the units are currently 5-10 meters apart in our lab.
1532e root bridge transmitter power = 20 (max)
1532i non-root transmitter power = 5 (minimum) ... on previous versions of the IOS I think this option was disabled on the non-root.
What I would do is get a baseline... connect the two PC's together and see what you speed is.... this is your basemine. Then connect the bridge and test again and make sure that the PC's are negotiating speed and duplex. I don't expect you to see real high throughput to be honest, I've deployed bridges and mesh and sometimes its hit or miss, but I have learned to accept the results. Code version also seem to have an affect, so there are other variables other than what the bridges negotiate at and what you are seeing... just keep in mind that the wireless link is half duplex also.
Thanks Scott, will try this.
BTW, just played around with the transmitter/client powers and managed to get a slightly better throughput.
root, transmit @ 20 (max), client @ 5 (min)
non-root, transmit @ 20 (max), client @ 5 (min)
Speeds are now 4.3 MB\s one way and 5.7 MB\s the other.
I need to leave for the day now but will be back in the morning to tinker some more.
The baseline of the 2 PC's between a single switch is 11.5 MB\s consistently so no problem there.
After more hours of tinkering we thought we would try something quickly and connected the 1532e to an autonamous 1142. After 10 mins of config we were seeing consistent speeds of 10MB\s !?!?
I'm really worried I've thrown away $6000 on kit that just isnt that compatible (1532e > 1532i) ... or has a buggy IOS/Hardware combination.
Any thoughts would be appreciated by a Cisco representative if at all possible as I doubt my supplier will take these back.
The baseline of the 2 PC's between a single switch is 11.5 MB\s consistently so no problem there.
11.5Mbps is pretty slow between two machines. Also when testing, I wouldn't have the radios powered up that high unless they are outside. Do you have these 1532's just laying on a table or are they mounted properly for testing.
Apologies, I may not have been clear. The figures quoted are actual throughput, so the PC's are connected to a switch (full duplex, 100mbps), time to transfer a 100mb file is approx 9 secs so seems fine.
What I usually do is set all the powers to the minimum when testing but the speeds were very poor so the internal environment may be a problem for these 1532 units as you say (our 1142 units run perfectly in the same environment).
The units are not mounted correctly just yet but we have done a rough test between the buildings we are connecting without much difference (the buildings are approx 100-200 meters apart). Today we are going to try the building > building connection again but use the 1532e > 1142 to see what throughput we get.
1532e and 1142 now setup in our lab (1142 is running an older autonomous IOS).
Seeing a throughput of upto 15 MB\s ! ... on 5Ghz - above 40hz (I think this is probably the maximum under 300Mbps n speed). Using a 1532i > 1142 shows similar throughput.
We cant get anywhere near this using the 1532e > 1532i which must point to problems with the IOS/Hardware compatibility.
I have a feeling that if we could run the older IOS on one or both 1532 units everything would be fine ??
Tried this but WPA1 wont work with EAP (it wants a shared key) so I would need to reconfigure a few things but wont have time until tomorrow. Having said that I think I recall that you can't run on 11n without using WPA2 and EAP.
Did you upgrade the Autonomous code?
Release 15.2.4-JB4 is the current version and they have a lot of resolved caveats since they released the 1530. I didn't see anything that popped out the same as your issue, but you never know...
How do we get our hands on this version ? Our version ...
|Top Assembly Serial Number:||FCZ1808H00G|
|System Software Filename:||ap1g3-k9w7-tar.152-4.JB3b|
|System Software Version:||15.2(4)JB3b|
Go to the downloads page and submit for entitlement. Since you are within 1 year of purchasing you should be able to get the updated version.
Well I emailed Cisco tech support and because I've not bought a contract with the kit they didnt really want to know and pointed me in the direction of my supplier. So trying to get an updated image from them now :(
We are seeing actual throughput speeds of 15MB\s with 1532 (e or i) > 1142 but less than half that using a 1532e>1532i. Very frustrated now.
Next option is to get rid of the kit and go with a different manufacturer I guess.
Thanks to everyone who tried to help out (especially TJ). We did get the updated IOS (JB4) but it made no difference. We have now given up on the 1532e>1532i setup but do have a working solution.
We have tested the 1532e>1142 on 20MHz between the buildings in question and get a fairly consistent 7-8 MB\s throughput. Once they are installed properly tomorrow we can start playing around with powers and frequencies.
I will leave this unanswered but I'd love to know if anyone has got a 1532e>1532i transferring at full 300 n speed (even the guys at Cisco!).
That is too bad..... Are you actually shooting the link through a window, since you are using an 1142?
This is now installed between the buildings (approx 150 meters apart) and is currently working at pretty much full speed on 5GHz/20MHz. Throughput is 11.5 MB\s both ways (same speeds we achieved when testing in the same room). Best we could get with the 1532e>1532i on 5GHz/20MHz was around 3 MB/s ! (40MHz gave us speeds between 4-6 MB/s)
Our users will be in the office shortly so time will tell with speeds and reliability.
Also worth noting that the 1532e>1142 was not happy on 2.4GHz. The throughput speeds were really bad like there is an IOS/Hardware issue with them communicating. Luckily 5GHz seems fine.
Link test between the 1532e/1142 bridge for anyone that is interested ..
|Link Test Output|
|GOOD (0 % retries)||Time||Strength(dBm)||SNR||SNR||Retries|
|Lost to Target|
|Lost to Source|
|Rates (Source/Target)||mcs15 100/100|
Just to add we are using 2x 3 meter Low Loss N-Type male to N-Type male cables (HDF400) with our 1532e. Apparently these have a loss of approx 0.36db per Meter for 5GHz applications.
I ran into the exact same problem. I could only move 24Mbps(3MBps) in either direction. I upgraded to ap1g3-k9w7-tar.153-3.JAB with absolutely no configuration change and now I can move 160Mbps in either direction. Sorry, I did not take note of the original version. It was the one that came from the factory. I am also set for 5Ghz, 40Mhz channel and below and default speed.
There is a bug CSCuo69578 (AP1532E/I bridge throughput is very low on 802.11n).
It has been solved in version 153-3.JAB.