Efficient tracing of failed nodes in sensor networks

In sensor networks, nodes commonly rely on each other to route messages to a base station. Although this practice conserves power it can obscure the cause of a measurement outage in a portion of the network. For example, when a base station ceases to receive measurements from a region of nodes it can't immediately determine whether this is because of the destruction of all the nodes in that region (due to an enemy attack, for example) or merely the result of the failure of a few nodes bearing much of the routing load. Previous solutions to this problem typically consist of re-running the route-discovery protocol, a process that can be quite expensive in terms of the number of messages that must be exchanged. We demonstrate that the topology of the network can be efficiently conveyed to the base station allowing for the quick tracing of the identities of the failed nodes with moderate communication overhead. Our algorithms work in conjunction with the existing functions of the network, requiring the nodes to send no additional messages.