DAPs send periodic reports to the DC, so it is easy for the DCto detect
when a DAP fails. In our current implementation, if the DC does not
receive any reports from a DAP for up to one minute, it flags the DAP
as a possible failure and does not assign any new clients to it. The
clients associated with the failed DAP get disconnected. These clients
immediately begin scanning for other DAPs in the vicinity by sending out
probe request messages. Other DAPs in the vicinity pick up these probe
messages and alert the central controller, which assigns these clients
to other DAPs, as per the association policy.
