#10397 NORM Not Tri: XOs may fail to communicate in noisy 802.11 environment

[Zarro Boogs per Child]

#10397: XOs may fail to communicate in noisy 802.11 environment
------------------------------------+---------------------------------------
           Reporter:  greenfeld     |       Owner:  dsaxena      
               Type:  defect        |      Status:  new          
           Priority:  normal        |   Milestone:  Not Triaged  
          Component:  connectivity  |     Version:  not specified
         Resolution:                |    Keywords:               
        Next_action:  never set     |    Verified:  0            
Deployment_affected:                |   Blockedby:               
           Blocking:                |  
------------------------------------+---------------------------------------
Changes (by wad):

 * cc: wad (added)
  * owner:  wad => dsaxena
  * component:  hardware => connectivity


Comment:

 The XO WLAN is not MIMO.  It switches between the antennas depending on
 which has the best signal.  Bear in mind Martin's recent report from
 Mexico, you may be seeing faulty antennas on some units.   In general the
 XO-1.5 WLAN units seem more even in quality than the XO-1 units, but time
 will tell.  On XO-1, the receive filters were such that in close
 situations (within a couple of meters) it was common to see a laptop
 receive broadcast packets sent on a non-overlapping channel.   The
 firmware was changed to filter these out, but that doesn't mean they
 aren't still stepping on valid traffic on that channel.

 We ran into this same problem with 1CC, hence my moving "into the country"
 for previous testing:
 http://wiki.laptop.org/go/Collaboration_Network_Testbed

 We previously considered shielding, etc.  You are correct, in that
 deployments rarely see the RF conditions at 1CC/11th fl. Miami.   If you
 need to know whether the problems you are seeing are due to the RF
 environment or the XO software, you will have to change RF environments.
 How big is your garage ?  I managed to fit fifty laptops into my
 basement...

 You are running into both fundamental problems with the communications
 stack on the XO (mainly collaboration, but also network) and with the
 nature of overloaded networks.   Once the network gets bad enough that a
 laptop can't establish a connection with the school server, that laptop
 will use salut instead of gabble and your wireless network just got much
 worse.

 TCP will just keep retrying until it gets through.  DNS (at the
 client/server level) uses TCP.   I found a WiFi spectrum analyzer
 invaluable in debugging these problems.  There were some days at 1CC when
 outside interference made any testing impossible.  Other days, it would be
 a broadcast storm generated by too many laptops running salut (and worse
 if they were in mesh mode!).

 Reassigning from hardware to connectivity.

-- 
Ticket URL: <http://dev.laptop.org/ticket/10397#comment:4>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system