tbs_mnp
asked on
Intermec CV41 roaming issues with Ruckus wireless infrastructure
Looking to the community for any experience and/or advice in regards to the following:
We have a Ruckus Wireless infrastructure utilizing all ZoneFlex 7372 AP's in a single building that is roughly 100,000sq-ft (manufacturing facility). The AP's are controlled by the Ruckus Virtualized Smartcell Gateway-vSCG (VM/Virtual appliance on VMware 5.5).
The AP's are on firmware 2.5.0.1.174
The vSCG is on version 2.5.0.1.174
We are only utilizing 2.4Ghz
Since the day this system was installed, we have experienced horrific connection & roaming behavior with our Intermec CV4’s ranging from application lag and application disconnects. The only application we use on the CV41 is Intermec's ITE (terminal emulation) software to establish a telnet session to our ERP system. The CV41’s are installed on mobile forklift trucks that travel throughout the facility at about 5-10mph on average.
We have been through months of investigation and modifications to both the Ruckus and Intermec systems. Intermec and Ruckus have both all been engaged and onsite to diagnose. The CV41’s have all been updated to their latest respective software levels specifically the Summit client/drivers. We have modified the Summit settings per Intermec and Ruckus recommendations (Ruckus Wireless Intermec Interoperability Guide-rev 1_14.pdf):
Aggressive Roam = on
Radio Mode = “BG rates full” (Only if the 5Ghz band will not be used.)
Power Save = “”CAM”
Tx Power = “Maximum”
Roam Trigger = “-70 dBm” (Set this option to work best for your environment)
Roam Delta = “5dBm”
Roam Period = “5 sec”
BG Channel Set = “1,6,11” (This is optional and should not be set if other channels are used.)
PMK Caching = “OPMK” (Set this if 802.1x is being used)
All recommended DTIM, beacon, etc. settings on the Ruckus side have been modified to match recommendations as well.
Has anyone even remotely had the same issues that we are experiencing?
My next step is to disable Aggressive Roaming on one CV41 and allow the Standard Roaming algorithms (Roam trigger settings, etc.) to take over.
Coverage is not an issue and kudos to Ruckus on that end. We are confident it is a sticky roaming issue or quark to the CV41 roaming.
Thank you all.
We have a Ruckus Wireless infrastructure utilizing all ZoneFlex 7372 AP's in a single building that is roughly 100,000sq-ft (manufacturing facility). The AP's are controlled by the Ruckus Virtualized Smartcell Gateway-vSCG (VM/Virtual appliance on VMware 5.5).
The AP's are on firmware 2.5.0.1.174
The vSCG is on version 2.5.0.1.174
We are only utilizing 2.4Ghz
Since the day this system was installed, we have experienced horrific connection & roaming behavior with our Intermec CV4’s ranging from application lag and application disconnects. The only application we use on the CV41 is Intermec's ITE (terminal emulation) software to establish a telnet session to our ERP system. The CV41’s are installed on mobile forklift trucks that travel throughout the facility at about 5-10mph on average.
We have been through months of investigation and modifications to both the Ruckus and Intermec systems. Intermec and Ruckus have both all been engaged and onsite to diagnose. The CV41’s have all been updated to their latest respective software levels specifically the Summit client/drivers. We have modified the Summit settings per Intermec and Ruckus recommendations (Ruckus Wireless Intermec Interoperability Guide-rev 1_14.pdf):
Aggressive Roam = on
Radio Mode = “BG rates full” (Only if the 5Ghz band will not be used.)
Power Save = “”CAM”
Tx Power = “Maximum”
Roam Trigger = “-70 dBm” (Set this option to work best for your environment)
Roam Delta = “5dBm”
Roam Period = “5 sec”
BG Channel Set = “1,6,11” (This is optional and should not be set if other channels are used.)
PMK Caching = “OPMK” (Set this if 802.1x is being used)
All recommended DTIM, beacon, etc. settings on the Ruckus side have been modified to match recommendations as well.
Has anyone even remotely had the same issues that we are experiencing?
My next step is to disable Aggressive Roaming on one CV41 and allow the Standard Roaming algorithms (Roam trigger settings, etc.) to take over.
Coverage is not an issue and kudos to Ruckus on that end. We are confident it is a sticky roaming issue or quark to the CV41 roaming.
Thank you all.
ASKER
Thank you craigbeck.
Something I forgot to mention. We did harcode the channels on every AP (1, 6, 11). There are 10 AP's in total throughout the plant.
Something I forgot to mention. We did harcode the channels on every AP (1, 6, 11). There are 10 AP's in total throughout the plant.
ASKER
craigbeck-
Would you modify the TX Power setting on the client (CV41) or the AP configuration?
Would you modify the TX Power setting on the client (CV41) or the AP configuration?
Using 1,6 and 11 is recommended as they're non-overlapping. You start to get problems when you have more than 3 APs though.
Consider a wide-open space. If you put 4 APs in a room, two of the APs must be on the same channel. Once you do that you need to make sure that these two APs can't 'see' eachother, otherwise they interfere with eachother. That's co-channel interference.
Sometimes we can use channels 1, 4, 8 and 11 (in North America, for example) in a high-density environment to minimize the effects of co-channel interference. You would therefore end up with no non-overlapping channels, so you'd get neighbouring-channel interference instead. If you're in Europe or Japan you could use 1, 5, 9 and 13 instead to get even less interference.
What I would say though is it's always better to manually set the channels you use if there's no interference from foreign networks or non-802.11 sources, whereas if you do have neighbouring networks, etc, you'd be better limiting which channels you use but allow the controller to dynamically choose the best channel for each AP.
You should ask for a passive survey to be performed in order to generate a heatmap for your AP coverage. That will show you if you're suffering from co-channel interference. If you are, consider reducing the TX power for each AP, or turn some off completely to test.
Consider a wide-open space. If you put 4 APs in a room, two of the APs must be on the same channel. Once you do that you need to make sure that these two APs can't 'see' eachother, otherwise they interfere with eachother. That's co-channel interference.
Sometimes we can use channels 1, 4, 8 and 11 (in North America, for example) in a high-density environment to minimize the effects of co-channel interference. You would therefore end up with no non-overlapping channels, so you'd get neighbouring-channel interference instead. If you're in Europe or Japan you could use 1, 5, 9 and 13 instead to get even less interference.
What I would say though is it's always better to manually set the channels you use if there's no interference from foreign networks or non-802.11 sources, whereas if you do have neighbouring networks, etc, you'd be better limiting which channels you use but allow the controller to dynamically choose the best channel for each AP.
You should ask for a passive survey to be performed in order to generate a heatmap for your AP coverage. That will show you if you're suffering from co-channel interference. If you are, consider reducing the TX power for each AP, or turn some off completely to test.
Would you modify the TX Power setting on the client (CV41) or the AP configuration?I'd modify the APs first (as I just said above).
The clients will generally transmit at their maximum, but that'll not cause near as much of a problem as the APs smashing their signal at full power.
ASKER
Understood craigbeck.
There is so much history with this porject that it is impossible to explain it all but a long story short: we had Ruckus and Intermec on site here. I have attached the current state AP placement and heat maps for reference. I thank you again for your insight. Do not let the AP placement map fool you, we only have 10 AP's in operation as noted in the document.
AP-Locations-Channels-MAC-Address.docx
Post-Installation-Testing.pdf
There is so much history with this porject that it is impossible to explain it all but a long story short: we had Ruckus and Intermec on site here. I have attached the current state AP placement and heat maps for reference. I thank you again for your insight. Do not let the AP placement map fool you, we only have 10 AP's in operation as noted in the document.
AP-Locations-Channels-MAC-Address.docx
Post-Installation-Testing.pdf
ASKER
Also, in your opinion (not to be held as gospel) could you explain and recommend a safe TX power level based on the third attachment to this thread.
TX-Power.bmp
TX-Power.bmp
Just looking through the docs you sent now, thanks.
Spotted the Canopy AP at the top of the plan. Canopy gear will interfere badly! It uses the same band, but it's not WiFi.
That's actually not a bad channel plan. I can't complain about it. The heatmap on the other hand... I'd say they've surveyed it too strictly. It looks like they've only gone as far as -72dBm. I'd go to -75dBm for a data-only deployment. That would give you a bit more space between the APs and that may alleviate the issues somewhat.
Why were AP34 and AP39 added? They're not on the heatmaps.
The TX power is never a safe figure. A lot rests on the pre-install RF survey. To produce a good survey you should use the same AP as you're going to deploy, with the same antenna(s). As well as that you should set the TX power of the AP to that of the lowest-powered client that you expect to use. This will do two things... 1) ensure that all clients can talk back to the AP in 100% of the coverage area, and 2) if you so suffer from a coverage hole (an AP dies, for example) the surrounding APs can power-up to fill the gap. That said, the APs should be set to the same TX-power once deployed, or dynamic power control should sort that for you.
Spotted the Canopy AP at the top of the plan. Canopy gear will interfere badly! It uses the same band, but it's not WiFi.
That's actually not a bad channel plan. I can't complain about it. The heatmap on the other hand... I'd say they've surveyed it too strictly. It looks like they've only gone as far as -72dBm. I'd go to -75dBm for a data-only deployment. That would give you a bit more space between the APs and that may alleviate the issues somewhat.
Why were AP34 and AP39 added? They're not on the heatmaps.
The TX power is never a safe figure. A lot rests on the pre-install RF survey. To produce a good survey you should use the same AP as you're going to deploy, with the same antenna(s). As well as that you should set the TX power of the AP to that of the lowest-powered client that you expect to use. This will do two things... 1) ensure that all clients can talk back to the AP in 100% of the coverage area, and 2) if you so suffer from a coverage hole (an AP dies, for example) the surrounding APs can power-up to fill the gap. That said, the APs should be set to the same TX-power once deployed, or dynamic power control should sort that for you.
ASKER
Spotted the Canopy AP at the top of the plan. Canopy gear will interfere badly! It uses the same band, but it's not WiFi.This is refering to a physical building roof canopy :)
Why were AP34 and AP39 added? They're not on the heatmaps.These are located in our front office, second floor and first floor respectively. They do not pertain to the mfg. shop floor as coverage bleed over is minimal due to the ocnstruction of the building.
Thank you again.
Ok cool that makes sense!
Have Ruckus done any kind of spectrum analysis?
Have you tried sticking a laptop on one of the vehicles to see if it suffers the same?
Have Ruckus done any kind of spectrum analysis?
Have you tried sticking a laptop on one of the vehicles to see if it suffers the same?
ASKER
Have Ruckus done any kind of spectrum analysis?I can not confirm or deny that. I honestly do not know.
Have you tried sticking a laptop on one of the vehicles to see if it suffers the same?This i have not tried but will keep it as a suggestion.
I modifed one CV41 this morning to turn aggressive roaming off and set the roam trigger to -80. i just got an email from the forklift driver stating his telnet session is locking up / lagging still.
I modifed one CV41 this morning to turn aggressive roaming off and set the roam trigger to -80. i just got an email from the forklift driver stating his telnet session is locking up / lagging still.Ok, I would definitely get a spectrum analysis done. It is possible that when the survey was done there was no noise/interference, but now there is.
ASKER
Understood. Thank you sir.
ASKER
One more question craig.
I have attached yet another drawing. This drawing displays the old Intermec WiFi infrastructure. No wireless devices in this facility are using this infrastructure anymore however the AP's are still on. Now, if you correlate the UHQ-UWN map to the other map outlining the Ruckus placements, you will see a number of the Intermec AP’s and active Ruckus AP’s are near each other with the same channel assignment. Would these cause interference/issues?
UHQ-UWN-Fiber-FloorLayout-CurrentState-r
I have attached yet another drawing. This drawing displays the old Intermec WiFi infrastructure. No wireless devices in this facility are using this infrastructure anymore however the AP's are still on. Now, if you correlate the UHQ-UWN map to the other map outlining the Ruckus placements, you will see a number of the Intermec AP’s and active Ruckus AP’s are near each other with the same channel assignment. Would these cause interference/issues?
UHQ-UWN-Fiber-FloorLayout-CurrentState-r
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Good idea! :-)
Too aggressive IMO. Set it to something lower, like -80dBm to test.
Ok, so...
If your coverage is good (ie, no coverage holes and plenty of overlap between APs), this may cause co-channel or neigbouring-channel interference issues which will cause degraded performance and/or disconnects.