Integration of heterogeneous sensor nodes by data stream management

Abstract

Wireless Sensor Networks (WSNs) will be an important streaming data source for many fields of surveillance in the near future, as the price of WSN technologies is diminishing rapidly, while processing power, sensing capability, and communication efficiency are growing steadily. Data-stream analyses should be distributed over the entire network in a way that the processing power is well utilized, the sensing is done in a semantically reasonable way, and communication is reduced to a minimumas it consumesmuch energy in general. Surveillance experts of different domains need technical experts in order to deploy those distributed data stream analyses. Data-stream queries often realize data-stream analyses. Especially surveillance scenarios that base on Sensor Data Fusion (SDF) will need the integration of heterogeneous data sources produced by potentially heterogeneous sensor nodes. This chapter overviews existing WSN middleware solutions, Stream Processing Systems (SPSs), and their integration. An approach that maps a global data-stream query to distributed and heterogeneous sensor nodes and SPSs opens a path to solve the problems mentioned above. Integration is achieved in two ways: semantic integration is done implicitly by the partitioning and mapping using rules that retain the semantics of the global query through the entire distribution and deployment process; technical integration is achieved during mapping and deployment with the help of the knowledge about platforms and connections. © 2010 Springer-Verlag Berlin Heidelberg.