Hi,
We are
actually building a small Mesh OpenThread network with a few devices (4
routers and 1 leader for example).
With
default parameters, if we stop the leader or reboot one router randomly,
the system takes several seconds or minutes to reconfigure itself. This is due
to the different default values specified in the Thread spec (MLE_END_DEVICE_TIMEOUT,
LEADER_TIMEOUT, ROUTER_SELECTION_JITTER etc..)
For the purpose of a demo, we
would like to have a system which ‘auto configure’ itself much more quickly,
lets say in a few seconds. For instance, if I power down the leader, I
want to see a new leader being elected as quickly as possible. If I stop a
router, I want the leader to be aware as fast as possible of the fact that the
router has been removed from the Mesh network. Etc..
Is it possible to have a set of
parameters and the API associated that we could use in order to highlight
the “auto reconfiguration” aspects of a typical Thread network and improve the
reactivity response time.
We can use functions like otThreadSetRouterSelectionJitter() and play with the poolperiod to improve reactivity.
Can you propose a set of values
that we could use for this App in order to highlight "Thread self reconfigurations" without breaking the overall robustness.
Thanks very much for your help.