Efficient Query Processing on Bulk Data of Sensor and Location

Currently, RFID tag is used for various applications. The environment of this paper is an existing application using semi-passive sensor tag and location device. Since this application needs watch, semi-passive sensor tag and location device to improve the transportation condition of trucks. These devices generate bulk data of sensor information and location information. However, processing this type of bulk data will generate heavy computation load by using existing stream data processor, which processes the real-time data stream of all registered queries are compared by all transferred bulk data. Furthermore, heavy computation causes drop of continuous query processing. To solve this problem, we propose two-level filtering technique. First level is to define MBB which is made by minimum value and maximum value of all bulk data for filtering the queries, so that the system does not need to process all the defined queries and select the candidate registered queries for the next step. Second level filtering is targeting at the bulk data and filtering by user defined query condition of registered query. And the system will process the selected query and generate result of continuous query of incoming bulk data. We also design and implement the architecture of our system, which has bulk data processing function.