Location-Centric Storage for On-Demand Warning in Sensor Networks

This paper presents a novel distributed data storage protocol addressing on-demand warning in sensor networks. Locations that store the record of an event are controlled by the distance to the event location and the intensity of the event. Higher intensity indicates that more sensors in larger geographic area should keep a copy of the event record. When a user passesby, a sensor can generate warning messages based on the records in its database. The closer to the event location, the more number of warning messages a user should get. This protocol scales well to large sensor networks and ensures that the storage load is well balanced. Performance analysis and simulation study indicate that our location-centric storage protocol can reach a stable state quickly and perform well when the event arrival rate follows a poisson distribution with the same mean λ.