Execution Of Continuous Query On Spatio-temporal Data Streams

At present, the location-based services (LBS) are in the stage of rapid development.In the location-aware environment, mobile devices generate large amounts ofspatio-temporal data, while servers are confronted with limited resources and strictrequirement of real time response. Traditional spatio-temporal databases whichmaterialize data in disk can not meet such demand. A new direction of research is tocombine the data stream query processors and spatio-temporal query processors.Designing a good algorithm to provide an efficiently query processing forspatio-temporal streams has become a hotspot.On the basis of analysis and research of existing data stream technology andspatio-temporal query technology, a algorithm of execution continuous queries onspatio-temporal data stream (CQST) is proposed. The CQST algorithm reducecomputational complexity by applying the spatio-temporal properties of moving objectsand queries efficiently to optimize queries execution:(1) In the stage of generatingquery plan，CQST introduces the concept of shared execution architecture to supportshared query processing by mapping different query logic to one spatio-temporalwindow operator.(2) A new tuple will be discarded or updated according to its“Expected Update Time”, which is calculate based on the position, velocity of movingobjects and queries at a time, so unnecessary spatial joins will be avoided.We expand open-source data stream system STREAM to support continuousspatio-temporal query. A spatio-temporal window operator supported CQST algorithmis designed and implemented in the system. Finally, algorithm is proved to be feasibleby performance testing and analysis.