Wireless Access

Hi Team.

One Question.

There is a tool PERMITTED LATENCY TESTING between the AP and the controller?

This tool must be characteristic of the controller.

Regards,

hi eparedesburga,

There is no tool available in the controller, but you can check with some of the following:

a>  ping <ip of ap>   (or from some desktop device, set to route via the controller, can do longer term ping)

b> show ap details advanced ap-name <the ap>

For b> above, here is some interpretation of some of the info (mostly in the first page of output)

(sg-3200) #show ap details advanced ap-name ap125

AP "ap125" Basic Information
----------------------------
Item                                                 Value
----                                                 -----
AP IP Address                                        192.168.1.9
<snip>
<snip>
ap RTT total, hiwmk:                                 10, 1
Currently in reglist                                 No
Reglist Entries, Exits:                              86 86
<snip>
<snip>

the line of interest here is "ap RTT tota, hiwmk".  In this case it has a value of 10,1 which means there has been 10 times where the RTT (in terms of response to a msg from controller to AP) has been between 1 and 1.99 seconds.

if the value showed 10,5 then it would imply that there was a moment where the latency was 5 seconds max, and there has been some total amount of latency of 10 seconds.

Also just below the above outpout, there are per message RTT counts, granulatiry is 1 second, so for 0-1.0 it will show as 0, for 1.01 to 2.00 it will show 1.

Rebootstraps and Control Messages Log
-------------------------------------
Recent Messages            Time now: Fri May  9 11:51:49 2014
---------------            ----------------------------------
Time Offset                Message details
-----------                ---------------
-286                       RCVD: KEEPALIVE len=20 peer=192.168.1.9 seq_num=602 rtt=0 result=OK
-885                       RCVD: KEEPALIVE len=20 peer=192.168.1.9 seq_num=601 rtt=0 result=OK
-1486                      RCVD: KEEPALIVE len=20 peer=192.168.1.9 seq_num=600 rtt=0 result=OK

hope that helps.

-jeff

slight clarification to

a>  ping <ip of ap>   (or from some desktop device, set to route via the controller, can do longer term ping)

if we are talking campus APs, yes you can route via the controller or can directly ping from some machine setup on a subnet near the controller. The closer to the controller the ping test platform then the more likely it is to approximate the network delay between controller and the APs.

Of course this will not add in any application delay (i.e. slow process on the controller or AP etc.). To check that, use the method outlined in b> as it cannot differentiate between network or application RTT.

For remote APs, which have two IP addresses (external IP and tunnel IP) then you have a choice - you can ping the external IP to get estimation of the network delay, and you can ping the internal IP (using the controller as the gateway) to get an idea if the controller or AP datapath is introducing any significant difference.  RAPs of course have a caveat that if they bootstrap and there is no static IP set for the inner IP, then they will always cycle to the next free IP in the address pool - have to be aware of that.

For long term ping monitoring, i would recommend a tool like smokeping (http://oss.oetiker.ch/smokeping/)