A method for removing duplicate data entries from a database that includes a plurality of application server nodes in an application server cluster includes tracking items using a plurality of ADA devices in a plurality of locations, receiving ADA observations over a network from the plurality of ADA devices, and balancing a load of data communications including the ADA observations so that the ADA observations are sent to different application server nodes of the application server cluster. The method further includes filtering duplicate observations at each application server node and separately filtering duplicate observations between the plurality of application server nodes.