Abstract: A system and method for distributing the fault management functions throughout a communications network. Each node in the network includes an associated fault agent/configuration agent pair. The fault agent receives alarm information and correlates the alarm information to identify a cause of a fault in the network. The fault agent then determines if correction of the fault can be handled at the current node. If so, the fault agent passes information identifying the fault to the associated configuration agent, which generates reconfiguration data for handling the fault and sends the reconfiguration data to subordinate nodes for implementation. If, on the other hand, the fault agent determines that the fault cannot be handled at the current node, the fault agent passes the alarm data, in summary form, to a higher level fault agent/configuration agent pair for handling at that higher level or for further processing and passing the data up the network until a level is reached at which the fault can be handled.