Wireless Access

last person joined: yesterday 

Access network design for branch, remote, outdoor, and campus locations with HPE Aruba Networking access points and mobility controllers.
Expand all | Collapse all

2.4Ghz Performance issues in Residence Halls

This thread has been viewed 0 times
  • 1.  2.4Ghz Performance issues in Residence Halls

    Posted Nov 15, 2012 01:50 PM

    The wireless network in our residence halls is installed with an older design:  APs in the hallway with downtilt antennas.  Devices in the buildings are experiencing moderate to severe performance issues on the network.  Devices connect to the controller and receive addresses via DHCP reliably.  There are a large number of interfering devices (both wifi and non-wifi).  Through both RF-Protect (demo) and AirMagnet I see an extremely high number of Block ACK DoS Attacks; after a discussion with TAC I believe these are not malicious, but a symptom of congested airspace.

     

    Network Details

    ArubaOS 6.1.3.4

    AP-105 or AP-124 (consistent within a building)

    Multiple styles of construction, from wood-framed to concrete block

     

    I came across the “Next Gen Architecture for Residence Halls” VRD and went through it.  The assumption in this document seems to be that a large number of devices are moving to 5Ghz radios.  Our clients are mostly low-end laptops, smartphones, and other consumer electronics; >90% of these devices connect at 2.4Ghz.  It does not seem to me that a complete redesign of our wireless installation to address 10% of the user base will significantly affect the performance issues that we are seeing.

     

    Summary:  I’ve been through the VRD for campus design (most suggestions implemented) and residence halls (skeptical that this will help at this time).  Does anyone have suggestions for providing good wireless service in residence halls with a large (and growing) population of 2.4Ghz devices?

     

    Thanks,

    John Pearson

    Wright State University

     



  • 2.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 15, 2012 02:17 PM

    Well thats really hard to tell..

    When you have a High Density Area you should have of course and you do dual band APS... and move the ones that can move to 5ghz to it... but as far i understood you, most are just 2.4ghz capable? is this true?

     

    You coul try using 802.1h/802.1d so the clients make less interference as they will negociate the output power.... the only bad thing with this isthat old clients with old drivers might not support this and may not see the wireless network...  You coul try but take that in mind....

     

    For the part of Wifi interference... are this wifi inside your network? because if they are you can detect them... if you got airwaves you can do a triangulation and see where are these rogue APs to turn them off...

     

    IF you have too many for example 105 in an area you can maybe turn off a few of those 2.4ghz radios so it will help with the interferece... or you can turn the feature of ARM Mode Aware If enabled, ARM will turn APs into Air Monitors (AMs) if it detects higher coverage levels than necessary.

     

     



  • 3.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 16, 2012 11:19 PM

    Hi John,

     

    I co-authored the Residence Halls paper and can provide some insight.

     

    The key point of the paper from an RF perspective is that hallway deployments do not work well, period.  Regardless of band. You can deliver significant additional capacity by moving to an in-room, denser deployment as explained in the paper.

     

    Now we do recommend 5GHz for this.  But you can do the same RF design with 2.4GHz only if that's appropriate for your environment.  You might intersperse a few dual-band APs to relieve congestion in the 2.4GHz band since you want to get 5GHz capable devices out of the way of the ones that don't have the option.

     

    With 3 channels and high wall attenuation, you'll get OK cell reuse.  The performance will be much better than what you have now because (i) the SNRs will be higher overall since you don't suffer hallway attenuation and (ii) the free space loss will be lower due to closer distance to AP.

     

    Higher SNRs will also improve resistance to external non-wifi interference.   You may also want to experiment with the "interferernce immunity" setting in the radio profile.

     

    Don't forget that hallway APs also interfere with each other!   They have a clear LOS to each other.   So if you are running ARM they are going to back down power which is the opposite of what you want.   

     



  • 4.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 18, 2012 11:33 AM

    Carlos, Chuck,

     

    Thanks for your insights.  There are some things here that we've considered/tried, and others we have not.  We haven't tried Chuck's in-room approach in the residence halls; they're "stick built", so room-to-room attenuation isn't terribly high.  Still, it seems like the most likely candidate to improve the situation.

     

    Regards,

     

    John



  • 5.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 21, 2012 07:29 PM

    Just to know

    1-what is the  is the procentage of packet loss if you do a ping for example to the gateway of your Wireless network? for example of 200 pings?

    2-This does not happen to clients connecting to 5ghz?

    3-Enet light of the 105 is on Green or Amber?

     

    Cheers

    Carlos



  • 6.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 26, 2012 02:01 PM

    Carlos,


    Thanks for stopping by again!

     

    Performance issues are somewhat fluid (as students migrate from the residence halls to campus and back).  Given that, I've pulled some representative data from the controller gui's dashboard:

     

    1) In problem areas, I typically see values like:

     

    frames dropped, AP to Client:  6% to 26%

    interference:  45% to 55%

    noise floor:  -75 dBm

     

    2) I have only diagnosed 1 problem area in 5Ghz on campus; there's an unknown piece of research equipment generating 80% non-wifi interference in 5Ghz.  That is not typical anywhere else on the network.

     

    3) Enet light on the 105s is Green (1Ghz).

     

    I have had a chance to speak with admins at other schools.  This type of behavior seems to be fairly typical.  I'm probably going to pursue two paths:  1) The design guide for Residence Halls and 2) some form of more direct rogue intervention (i.e. knocking on doors).

     

    Have a great day,

    John



  • 7.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 26, 2012 02:09 PM

    Your problem is defenily non 802.11 interference...

     

    Channel Noise Floor is above 80 its okay but  If Noise Floor is lower then too much non-802.11 interference, which is your case...

     

    I remenber that value its a documentation of Aruba of troubleshooting...

    Also i remenber of that document that the Channel Frame Retry Rate % should be below 30% if is higher then there is too much interference..

     

    Do you have RF protect license? you can put some APS on Spectrum analysis mode to see what is the source of interferenfe that will help you to find what is it...

     

    Anyways

    Frame dropped is too high....i mean more than 10% on packet drop is too high i even consider more than 5% is too high.. on packet drops on a ping test

     

    can you check the frames retry?

    do it with this command.

    show ap active ip-addr <ap ip address>

     



  • 8.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 26, 2012 02:12 PM

    Check if the frame retry is higher than 30%

    You will need to check on some APS...

     

    Also

    How many clietnts you got for each AP connnected at the same time?

    Is not recommended for at least AP 93 no more than 30 for good performance... or if its ap 105 no more than 30 per band...

     

    Do you have too many users on one AP on the same band?? at the same time?



  • 9.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 26, 2012 03:20 PM

     

    A noise floor value of -75dBm would be a serious cause for concern.  

     

    Getting either a portable or an AP-based spectrum analyzer in there is the next step for sure.

     

    Once you have spectrum captures, please share back and we can help analyze.



  • 10.  RE: 2.4Ghz Performance issues in Residence Halls

    Posted Nov 29, 2012 10:41 AM

     I would suggest planning for a large increase in 5GHz traffic. One of the factors which pushed most vendors to 2.4GHz is no longer there (cost of chips). Instead, looking ahead we see a requirement for compatibility with 802.11ac which is 5GHz only. I suspect you will see a growing trend of 5GHz kit very soon. (iPad 3 + iPhone 5 for starters!)