Query-based Wireless Sensor Storage Management for Real-time Applications

In this paper we propose location aided storage management for data centric storage in wireless sensor networks. Sensor networks store data in three places: external, in-network, local. In-network storage allows sensor nodes as storage nodes for sensed data in the sensor network. This form in-network storage in sensor networks is called data centric storage and is achieved by geographic hashing. Geographic hashing refers to hashing of data types to geographic locations. These hashed locations are mapped to nearest sensor nodes as storage points. One interesting observation in geographic hashing is the hashed locations are mapped to static storage nodes, and the mapping is not dependent on query traffic information. If queries are originating from far areas from the given storage node then the amount of traffic generated is high. The traffic can be decreased if the storage nodes are nearer to the query points. This observation leads us to propose a location aided storage management in which storage nodes are shifted towards the query regions. Our contribution from the traffic improvement point of view is ´shifting of storage nodes towards the query regions with new query region and storage node formation algorithms´. The improvements are decrease in query traffic and faster retrieval times. We validate the proposed methods by simulations.