Controllerless Networks

I am having an issue with (2) IAP clusters where the wireless download speeds are no 1/3 of wired speed but upload speeds are very close to wired speed.  At other locations with the same equipment we are getting much better download performance and just not sure where to start checking in the configs of the switches (Cisco) and/or IAP's.  Any advice would be greatly appreciated.

Try turning the power down on the AP a little and see if there is a difference.

Thanks for your reply.  Attached are what my current power settings are.  What level would you suggest going to?

I would try making the max 18 on both bands.  It says it is transmitting at 22/23 right now.  Asymetric speeds can be traced back to asymetric power.

I backed down the power and the issue still exists.  I am getting high 80's to 90 mbps upload from a client but 20-30 mbps download.

What band is your client connecting on?  What type of client is it?  What are you using for a speedtest?

Your last screenshot had both of your APs on channel 100E.

Do you have "Broadcast filter all" enabled on your SSID?

Adaptive Radio Management is set so I'm not quite sure why it has selected the same channel in a smaller building like it has.  I have broadcast filtering off for that SSID (see attached).  I am using a phone that gets good speeds everywhere on 5 GHz.  Same results on laptops and such in this building.

Here's another wierd thing.  The 2.4 GHz now has 2 radios on the same channel (see attached)

---

aruba@qsd.wednet.edu wrote:

Here's another wierd thing.  The 2.4 GHz now has 2 radios on the same channel (see attached)

---

Not wierd.  Your AP had 1/6/11 before, but channel 6 might have more interference than putting the other AP on channel 11.

You should have broadcast filter ALL or ARP enabled for the best performance.  

APs come up in channel 100 at first, so it is possible that because you have client aware checked, it did not change channels while clients are connected.

It is also entirely possible that your client is too close to the AP, so that itself could cause distortion and reduced performance.  Make sure the client being tested is at least 3 meters away.

- Uncheck Client Aware

- Change Broadcast Filtering to ALL or ARP.

- Retest

I have set Broadcast Filtering to All, disabled Client Aware and nothing changed.  Rebooted AP02 and it is still on the same channels that the other AP's are on.  See attached.

Ok.

Rebooting an AP does not speed up changing the channel.  Only time can, really.

I would try unchecking 80 mhz channels and running your test again.

There are a myriad of reasons why you would be having asymetrical speeds.  What speedtest are you using?

Same results after unchecking 80 MHz.  I am currently using Ookla speed test and realize there can be variation there.  This problem at this location has been ongoing for sometime and I've tested in the past with iPerf back to a server of ours with the same type of results.

Okay.  It could very well be an RF issue, then, since it is site-specific and it is not happening anywhere else.

I would tend to agree with that theory, however I'm just not sure where to go.  Here are some snapshots of what looks like higher interference levels than at most of our other schools.

That is the number of foreign APs that are seen but it does not draw a direct line to interference.

I would check the channel utilization for each AP at your site.

Attached is my channel utilization. 

That does not seem to be too high.  What version of InstantOS is this?

6.4.4.8-4.2.4.4_57496 But older versions behaved the same.

If you are only having problems at that location, it could be something at that location, network-wise, or a combination of things at that location.

I don't believe it is the switches setup at that location that is directly the issue because I can place a hard wired client on one of the "wireless" vlans in question and the throughput is perfect.  Could I send you the the setup of the cisco's interface that connect to the IAP and see if you can spot any issue?

Also, to prove that it isn't the network (Cisco side of things) I could setup a iPerf server hard wired to the switch and then do a speed test from a hard wired client, all on the same VLAN as the SSID.  Would that be a good proof of concept?  Then test that same server against a wireless client?  Then we are down to AP config or RF, right?

Here is what my Cisco Switches Interface is setup as -

description WAP
switchport trunk encapsulation dot1q
switchport trunk native vlan 250
switchport trunk allowed vlan 3,48,201,208,250
switchport mode trunk

Attached is the wired network profile.  Hopefully you see something out of wack here.

Well, I might see one issue.  I didn't initially setup the wireless profiles of the AP's, the reseller did and now I'm seeing that they have the Native VLAN in that profile being VLAN1 and I have the Cisco side being VLAN250.  Not even sure how these units could work if I had a Native VLAN mismatch.  Thoughts?

I don't know if that would affect throughput.  The native VLAN is simply the untagged VLAN.  It should not really affect throughput.

I think you should run an iperf3 test from the IAP to an iperf server to rule out a wired issue between the IAP and your network.

http://www.arubanetworks.com/techdocs/Instant_423_WebHelp/InstantWebHelp.htm#CLI_commands/speed-test priv.htm

I didn't even know there was a built in utility for iPerf on an AP.  Thank you, this will help troubleshooting a bunch!  I will let you know the results in a bit...probably in the next hour or so.  I truly appreciate you time!

There doesn't look to be a wired issue.  iPerf server hardwired and communicating with the AP gives results in the high 80's to low 90's consistently.  Put the server on the wireless (5 GHz, showing 400 mbps connection) and here are some of the results which are very spotty -

Accepted connection from 10.1.250.5, port 63239
[  6] local 10.2.48.3 port 49221 connected to 10.1.250.5 port 63240
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-1.00   sec   256 KBytes  2.09 Mbits/sec
[  6]   1.00-2.00   sec   640 KBytes  5.24 Mbits/sec
[  6]   2.00-3.00   sec  6.00 MBytes  50.3 Mbits/sec
[  6]   3.00-4.01   sec  10.8 MBytes  90.2 Mbits/sec
[  6]   4.01-5.00   sec  5.88 MBytes  49.4 Mbits/sec
[  6]   5.00-6.00   sec  10.4 MBytes  87.2 Mbits/sec
[  6]   6.00-7.00   sec  11.0 MBytes  92.2 Mbits/sec
[  6]   7.00-8.00   sec  7.50 MBytes  63.0 Mbits/sec
[  6]   8.00-9.01   sec  6.12 MBytes  51.1 Mbits/sec
[  6]   9.01-10.01  sec  1.75 MBytes  14.7 Mbits/sec
[  6]  10.01-11.01  sec  1.75 MBytes  14.7 Mbits/sec
[  6]  11.01-12.01  sec  2.88 MBytes  24.1 Mbits/sec
[  6]  12.01-13.01  sec  9.00 MBytes  75.5 Mbits/sec
[  6]  13.01-14.01  sec  9.75 MBytes  81.8 Mbits/sec
[  6]  14.01-15.01  sec  10.9 MBytes  91.2 Mbits/sec
[  6]  15.01-16.00  sec  11.0 MBytes  92.8 Mbits/sec
[  6]  16.00-17.01  sec  10.4 MBytes  86.5 Mbits/sec
[  6]  17.01-18.01  sec  11.0 MBytes  92.2 Mbits/sec
[  6]  18.01-19.01  sec  9.88 MBytes  82.9 Mbits/sec
[  6]  19.01-20.01  sec  2.50 MBytes  21.0 Mbits/sec
[  6]  20.01-20.42  sec  0.00 Bytes  0.00 bits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-20.42  sec   139 MBytes  57.2 Mbits/sec                  sender
[  6]   0.00-20.42  sec  0.00 Bytes  0.00 bits/sec                  receiver
-----------------------------------------------------------
Server listening on 49221
-----------------------------------------------------------
Accepted connection from 10.1.250.5, port 63241
[  6] local 10.2.48.3 port 49221 connected to 10.1.250.5 port 63242
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-1.00   sec   378 KBytes  3.09 Mbits/sec
[  6]   1.00-2.00   sec   190 KBytes  1.56 Mbits/sec
[  6]   2.00-3.00   sec  14.6 KBytes   120 Kbits/sec
[  6]   3.00-4.01   sec   334 KBytes  2.72 Mbits/sec
[  6]   4.01-5.00   sec  3.96 MBytes  33.5 Mbits/sec
[  6]   5.00-6.00   sec  10.0 MBytes  84.0 Mbits/sec
[  6]   6.00-7.00   sec  9.73 MBytes  81.7 Mbits/sec
[  6]   7.00-8.00   sec  10.6 MBytes  89.1 Mbits/sec
[  6]   8.00-9.00   sec  10.1 MBytes  84.5 Mbits/sec
[  6]   9.00-10.00  sec  9.48 MBytes  79.6 Mbits/sec
[  6]  10.00-11.01  sec  8.68 MBytes  72.3 Mbits/sec
[  6]  11.01-12.01  sec   954 KBytes  7.82 Mbits/sec
[  6]  12.01-13.00  sec   644 KBytes  5.31 Mbits/sec
[  6]  13.00-14.00  sec   878 KBytes  7.19 Mbits/sec
[  6]  14.00-15.00  sec  5.51 MBytes  46.2 Mbits/sec
[  6]  15.00-16.00  sec  9.10 MBytes  76.4 Mbits/sec
[  6]  16.00-17.00  sec  9.51 MBytes  79.8 Mbits/sec
[  6]  17.00-18.00  sec  9.94 MBytes  83.4 Mbits/sec
[  6]  18.00-19.00  sec  10.2 MBytes  85.9 Mbits/sec
[  6]  19.00-20.00  sec  10.3 MBytes  86.2 Mbits/sec
[  6]  20.00-20.35  sec  2.77 MBytes  66.6 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-20.35  sec  0.00 Bytes  0.00 bits/sec                  sender
[  6]   0.00-20.35  sec   123 MBytes  50.8 Mbits/sec                  receiver
-----------------------------------------------------------
Server listening on 49221
-----------------------------------------------------------
Accepted connection from 10.1.250.5, port 63243
[  6] local 10.2.48.3 port 49221 connected to 10.1.250.5 port 63244
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-1.01   sec   256 KBytes  2.07 Mbits/sec
[  6]   1.01-2.01   sec  4.50 MBytes  37.9 Mbits/sec
[  6]   2.01-3.01   sec  7.12 MBytes  59.8 Mbits/sec
[  6]   3.01-4.01   sec  8.88 MBytes  74.7 Mbits/sec
[  6]   4.01-5.00   sec  9.38 MBytes  78.9 Mbits/sec
[  6]   5.00-6.01   sec  10.9 MBytes  90.7 Mbits/sec
[  6]   6.01-7.01   sec  10.8 MBytes  90.2 Mbits/sec
[  6]   7.01-8.01   sec  9.12 MBytes  76.2 Mbits/sec
[  6]   8.01-9.01   sec   896 KBytes  7.34 Mbits/sec
[  6]   9.01-10.00  sec  1.88 MBytes  15.9 Mbits/sec
[  6]  10.00-11.01  sec  1.62 MBytes  13.5 Mbits/sec
[  6]  11.01-12.01  sec  9.25 MBytes  78.0 Mbits/sec
[  6]  12.01-13.00  sec  10.6 MBytes  89.8 Mbits/sec
[  6]  13.00-14.00  sec  9.88 MBytes  82.8 Mbits/sec
[  6]  14.00-15.00  sec  11.0 MBytes  92.0 Mbits/sec
[  6]  15.00-16.00  sec  10.5 MBytes  88.1 Mbits/sec
[  6]  16.00-17.01  sec  10.9 MBytes  91.0 Mbits/sec
[  6]  17.01-18.01  sec  9.38 MBytes  78.6 Mbits/sec
[  6]  18.01-19.01  sec  2.25 MBytes  18.9 Mbits/sec
[  6]  19.01-20.01  sec  1.25 MBytes  10.5 Mbits/sec
[  6]  20.01-20.04  sec  0.00 Bytes  0.00 bits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-20.04  sec   140 MBytes  58.7 Mbits/sec                  sender
[  6]   0.00-20.04  sec  0.00 Bytes  0.00 bits/sec                  receiver
-----------------------------------------------------------
Server listening on 49221
-----------------------------------------------------------
Accepted connection from 10.1.250.5, port 63245
[  6] local 10.2.48.3 port 49221 connected to 10.1.250.5 port 63246
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-1.01   sec   233 KBytes  1.89 Mbits/sec
[  6]   1.01-2.01   sec   378 KBytes  3.10 Mbits/sec
[  6]   2.01-3.01   sec   235 KBytes  1.93 Mbits/sec
[  6]   3.01-4.00   sec  2.61 MBytes  22.0 Mbits/sec
[  6]   4.00-5.00   sec  7.68 MBytes  64.5 Mbits/sec
[  6]   5.00-6.00   sec  10.3 MBytes  86.7 Mbits/sec
[  6]   6.00-7.00   sec  8.90 MBytes  74.7 Mbits/sec
[  6]   7.00-8.00   sec  10.4 MBytes  87.5 Mbits/sec
[  6]   8.00-9.00   sec  10.7 MBytes  89.7 Mbits/sec
[  6]   9.00-10.00  sec  10.5 MBytes  88.4 Mbits/sec
[  6]  10.00-11.02  sec  1.76 MBytes  14.5 Mbits/sec
[  6]  11.02-12.02  sec  1.04 MBytes  8.71 Mbits/sec
[  6]  12.02-13.02  sec   953 KBytes  7.81 Mbits/sec
[  6]  13.02-14.00  sec  3.26 MBytes  27.8 Mbits/sec
[  6]  14.00-15.00  sec  8.30 MBytes  69.6 Mbits/sec
[  6]  15.00-16.00  sec  10.1 MBytes  84.6 Mbits/sec
[  6]  16.00-17.00  sec  10.1 MBytes  84.9 Mbits/sec
[  6]  17.00-18.00  sec  9.86 MBytes  82.7 Mbits/sec
[  6]  18.00-19.00  sec  9.80 MBytes  82.2 Mbits/sec
[  6]  19.00-20.00  sec  10.1 MBytes  84.5 Mbits/sec
[  6]  20.00-21.01  sec  1.38 MBytes  11.5 Mbits/sec
[  6]  21.01-21.09  sec  56.1 KBytes  5.57 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  6]   0.00-21.09  sec  0.00 Bytes  0.00 bits/sec                  sender
[  6]   0.00-21.09  sec   129 MBytes  51.2 Mbits/sec                  receiver
-----------------------------------------------------------

WIth that information do you have any ideas where to go from here or additional types of test I should preform to narrow down the problem?

You should open a TAC case to get to the bottom of this.  They will be able to collect your logs and tell you what they think.  I am just guessing with limited information here.

Fair enough.  Is that just the regular support #?  I only ask because I've done this dance with them on this topic once before and didn't get very far at all.  These types of things can be tough to tackle remotely.

If at any point you feel like you are not getting anywhere with TAC, you can ask to have the case escalated.  In all fairness, when problems are brought to TAC, they have to figure out how to completely replicate your environment to theirs to even have a clue what is wrong.  There are many combinations of ways to break things; some that have been observed, and others that have never been observed.  You mentioned that you have sites that do not have your issue.  Try to figure out if there are any differences and bring that to TAC's attention, if possible.

I haven't got around to calling TAC yet, however we had a no school day due to weather the other day and decided to do a few more speed tests and guess what...the speeds were what I would expect!  So we have this location and another school on the same private VLAN provided by our local PUD.  Decided to start working at the location in question and started adding devices to the wireless to see when it would start to choke.  Opened a Chromebook or 2.  Same, good speed.  Opened a Windows laptop and boom.  The speeds died down.  Looked at the task manager of the laptop and Windows updates were happening and were hogging a TON of bandwidth.  My guess is that if ANY device jumps onto this private VLAN and starts downloading Windows updates on that IAP cluster it just causes congestion and VERY poor speeds.  What is the suggest method to mitigating this?  We use a WSUS server but that is at a remote location from these (2) sites on the PUD VLAN.  I have disable WUDO via GPO on our windows domain clients (including laptops) but I have not throttle BITS via GPO which might be a thought.  I would prefer to do this through the wireless but I don't want to throttle other traffic per se.   Let me know how you would suggest tackling this.  Thanks again for all your help and putting up with me.  There are just so many variables that are tough to eliminate but luckily we had a snow day!

I should also mention that I have AirWave so that might be a piece to the puzzle of solving QoS when it comes to Windows Updates?

I had a similar problem and solved it by unchecking "legacy only" in RF settings/ radio