B/G Network Latency Issue

[Cisco1619]

I am having an issue with my B/G networks in my environment and I can't figure out why. When I run a continuous ping on clients connected to my B/G radio networks, there is a lot of variations in the ping responses and frequent dropped pings. One the same PC and location, if I connect to my A radio network, I get good response times with no pings. I can't find anything in the configurations to account for this, so I was wondering if anyone might have something that could spark a new lead to check. Below is an example of some of my ping responses.

A Radio

Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=2ms TTL=128
Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=4ms TTL=128
Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=3ms TTL=128
Reply from 10.203.33.254: bytes=32 time=4ms TTL=128
Reply from 10.203.33.254: bytes=32 time=23ms TTL=12
Reply from 10.203.33.254: bytes=32 time=2ms TTL=128
Reply from 10.203.33.254: bytes=32 time=53ms TTL=12
Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=1ms TTL=128
Reply from 10.203.33.254: bytes=32 time=2ms TTL=128
Reply from 10.203.33.254: bytes=32 time=2ms TTL=128

B/G Radio

Reply from 10.204.128.125: bytes=32 time=61ms TTL=122
Reply from 10.204.128.125: bytes=32 time=307ms TTL=122
Reply from 10.204.128.125: bytes=32 time=47ms TTL=122
Reply from 10.204.128.125: bytes=32 time=82ms TTL=122
Reply from 10.204.128.125: bytes=32 time=127ms TTL=122
Reply from 10.204.128.125: bytes=32 time=171ms TTL=122
Reply from 10.204.128.125: bytes=32 time=225ms TTL=122
Reply from 10.204.128.125: bytes=32 time=47ms TTL=122
Reply from 10.204.128.125: bytes=32 time=176ms TTL=122
Reply from 10.204.128.125: bytes=32 time=12ms TTL=122
Reply from 10.204.128.125: bytes=32 time=54ms TTL=122
Reply from 10.204.128.125: bytes=32 time=97ms TTL=122
Reply from 10.204.128.125: bytes=32 time=47ms TTL=122
Reply from 10.204.128.125: bytes=32 time=188ms TTL=122
Reply from 10.204.128.125: bytes=32 time=16ms TTL=122
Reply from 10.204.128.125: bytes=32 time=167ms TTL=122
Reply from 10.204.128.125: bytes=32 time=213ms TTL=122
Reply from 10.204.128.125: bytes=32 time=43ms TTL=122
Reply from 10.204.128.125: bytes=32 time=95ms TTL=122
Request timed out.
Reply from 10.204.128.125: bytes=32 time=3ms TTL=122
Reply from 10.204.128.125: bytes=32 time=3ms TTL=122
Reply from 10.204.128.125: bytes=32 time=5ms TTL=122

These response times are a little better than what I see on the floor. On my production floor, I see times in the 1000ms - 2000ms responses mixed in with these and more drops. One of the built in tools on my handheld scanner that uses this SSID says that the quality is 98%-100% consistently with the RSSI at -65 ~ -72. I have a WiSM2 inserted in my 6509 MDF as a controller. I am using Cisco 3502i and Cisco 3502e with AIR-ANT1728 antennas attached to them spaced at 3.5" and AIR-ANT5160V-R also spaced at 3.5". Originally my contractor was to space these at 5" but do to some miscommunication, they placed them closer. They are going to go through and space them at the 5" spacing, but I am not sure if that will solve the issues or not. The response times are marginally better when I am on the 3502i APs in my office.

 Any input from the group would be greatly appreciated. I have been working on this for weeks, but I can't find any thing to explain these symptoms.

[Freerk Terpstra]

What is the output of a "show client detail MAC" in both cases? You are saying that you are having this problem with multiple clients, are those the same type of clients? What is the load on the 2.4Ghz radio's? Also, is there a possibility that you fire-up a spectrum analyzer to see how the 2.4Ghz spectrum looks like?

[Cisco1619]

Sorry for taking so long to respond. This issue has had me on the floor a lot lately, and I have not been able to get back on here. We finally were able to get it under control by putting the 1Mbps and 2Mbps data rates on the 2.4Ghz channel to "disabled". This was the recommendation by the scanner manufacturer. It appears to have worked very well. We also upgraded out controller code from 7.0.220.0 to 7.4.121.0. So, at the moment, we seem to be running much better. Cisco did recommend that I turn off a few of the 2.4Ghz APs since our deployment is very dense for the 2.4Ghz range. They said that might be causing some of the issues. I am planning on trying this over the next few weeks. I appreciate all of the input on this issue!

[Salma Sulthana]

Hi,

Thanks for the update, yes, it's always better to disable 1 and 2 Mbps and keeping 5.5 and 11 Mbps as mandatory. some of the scanner will have issues when you keep Aironet IE enabled under WLAN. It's good that you upgraded software to 7.4.121.0 it's an MD release and please do not upgrade to any  software if you are not hitting Bug.

Regards

Salma

[Scott Fella]

This is were testing helps a lot.  Since you can disable 1mbps and 2mbps, and the scanners and other devices still work. You can also try to disable 5.5mbps and set 11mbps as mandatory.  Some b/g devices still require 802.11b mandatory rates enabled, so this is where testing helps.  When b/g devices don't require an 802.11b mandatory rate, well you can either disable all 80211b rates and set one or two of the 802.11g rates as mandatory, but you need to know how your rf is on the floors.  If you have AP's in rooms/offices along with the hallway's, for example and you also know that you have more than enough coverage, you can set the mandatory rate to 24mbps for example and support the other rates.  Also tweaking the TX power helps, by setting the max TX and min TX on the global setting in the WLC.  I have never had to disable and 2.4ghz radios even in very high density deployments, because you can tweak the TX power and set the data rates.  You would do this even for the 5ghz to be honest.  

Scott

[Salma Sulthana]

Hi,

As you have issues with the APs using external antennas, it's very important that you mount the APs correctly by doing a proper site survey.

as you know, Radio A has better speed and less interference with more number of non overlapping channels compared to radio B/G, the response time will be fast on Radio A.

There are too many delays in the pings, make sure that there is no co-channel interference and the antenna gain is correctly updated on the external antenna APs.

However, it will be good if you can collect packet captures.

Regards

Salma Sulthana

[Scott Fella]

If you are seeing this everywhere with the same laptop, then usually an upgrade to the client drivers are required.  Especially if you upgraded the WLC to a newer code from a very old code.... Seen this happen many times. 

If the issue is in certain locations of the building, then it can be interference or high channel utilization in the area. Even if you have signal, you need to also look at what the SNR is and the channel utilization.  

You have only three non overlapping channels on the 2.4ghz and more available channels on the 5ghz. You stating that the 5ghz is fine in that location, but the 2.4ghz is bad, well from my experience, it does seem like it's either interference or high channel utilization and low SNR on the 2.4ghz in that area.  You would have to try to eliminate the interference if you can and maybe review the power levels and data rates if your using RRM.  

Scott