Design And Implement Of The Semantic Sensor Data Management System Based On HBase

Ontology is used to annotate sensor web data so as to enhance its semantic in Semantic Sensor Web. The Semantic Sensor Web takes use of RDF to store data, and it is beneficial to data management and data interconnection. With the development of Semantic Sensor Web, massive RDF dataset have been produced. Semantic sensor web data has the features of spatial and temporal and massive. So semantic sensor web is facing a big challenge now, which is the management and application of semantic sensor data, especially the storage and query of semantic sensor data.In order to implement the storage and query module of semantic sensor data, the thesis proposes a system structure, which semantic sensor data by form of RDF can be stored on Hadoop by MapReduce and HBase. This system consists of sensor network layer, persistent layer, Hadoop layer, function layer and application layer. The function layer mainly consists of storage module and query module. In order to implement the storage function of semantic sensor data by form of RDF, this thesis uploads stream RDF dataset to Hadoop file system, then this file in Hadoop file system is uploaded to HBase table by use of MapReduce and HBase API. In order to implement the query function of of semantic sensor data by form of RDF, this thesis designs the method of parsing query statement, task distribution method, continuous query method by use of thread, data query method by use of HBase API and tree structure, solution sequence processing method, Aggregate function method and form processing method, which implements the basic semantic query of SPARQL and the continuous semantic query.The thesis designs management system based on the architecture of HDSW and implements the storage and query of semantic sensor data by form of RDF. Finally, experiments are designed for storage and query method. The storage method experiments show the storage method of HDSW has the feature of scalability, efficiency and low space efficiency. The query method experiments prove the query method of HDSW has the feature of scalability and efficiency. So the HDSW system promotes the management and application of semantic sensor data to develop.