gfps/4.1+Observed+standard+deviation+is+too+low,+.........md

4.1 Observed standard deviation is too low, ........

BB2BUG-17449 - [BB2][RTL8763EF/ROM][GMI][沃莱特][Harman][T520]: Finder测项中，finder_advertisement rotation测试失败 Closed

Hi Avner:

We test advertising rotation failed, could you help see the reason, thanks

\

Hi Danni,

The reason that this test fails is that the device under test rotates EID/MAC pretty much precisely every 1024 seconds (standard deviation is 0.83 seconds).

Per spec the exact rotation time should be randomized, please refer to the "ID Rotation" section.

\

Hi Mark:

1.Does the positive randomized time factor have a value range? Can you provide some reference ranges?

2.Is the standard deviation must be greater than 0.83?

\

+David Lazarov Can you please attach the latest version of the spec which includes the recommended randomization values?

 

Hi Danni,

 

1. The recommended size of the randomization window is 20% of the interval, i.e. 204 seconds. So when the device needs to compute the next time point to rotate EID/MAC, it's going to be *floor((last_rotation_time_seconds + 1024) / 1024) * 1024 + random(0, 204). *
2. Standard deviation of 0.83 seconds practically means there's no randomization. We expect standard deviation to be much higher, typically 50-60 seconds, or more.

\