Forum

Hi,

Can anyone suggest me what would be a good way to test roaming?

The scenario is that APs are going to be connected to a controller and will need to simulate a train speed of around 100km, roaming between APs, transferring video feed.

How can we do this in a lab situation?

Any suggestion will be very much appreciated

1. Using test CLIs in the APs to control power.
- Frequency of power change, simulates some amount of speed.
2. Usage of tools like veriwave, ixwlan to simulate client(if client is the device that is roaming).
3. Using rf attenuators to vary power.
4. Creating a test environment outdoors and driving through it in a vehicle.
5. Protocol level checking would surely eliminate 95% of issues.

Can you specify what product are you using? Is the device in the train wireless station or AP?

I have recently done a similar work.

Thanks for your help.

The WLAN system will be Aruba based APs and controller however the system inside the train can be thought of as a client rather than an AP.

I am not quite clear how your first few points can simulate train running. Can you please elaborate a bit more?

Thanks once again

When does a client decide to roam?
- When it finds a better AP.
- RSSI is the most used parameter for roaming.
- When a client sees that its old AP has a lower RSSI than the new one (usually greater than a threshold, say of 10dBm), it decides to roam to the new AP.

This can be simulated by changing the power of AP or by dynamically adding attenuation.

If you want to see what is happing, you can use OmniPeek with multiple RF grabbers set on each channel (assuming 1,6,11) and see the packets as the handoffs occur.