Research On The Technology Of Continuous Query Processing Over Data Stream

The development of the internet and communication equipment has enabled new applications which include: financial analysis, network monitoring, telecommunication data management, sensor networks, etc. In these applications, the data which we called data stream is multidimensional, continuous, rapid, and time-varying, which should not only meet the needs of traditional ad-hoc query, but also the continuous query for special. In the aspect of the data assumptions and the query processing technology, the traditional database management systems can not satisfy these applications. As a result, the technology of efficiently continuous query processing for these applications, however, became one of the hottest parts in database research.In order to reach the demands mentioned above, our purpose is to do some research on the technology of continuous query processing over data stream and the data stream management system. The main contents done in this paper are as follows. First, we extend the data model of data stream. Second, we design a new query language named SKQL. Third, an approach is proposed to store and query historical stream data by using multi-layer recursive sampling method and H-T structure, which indexes the aggregation of historical stream data and supports all kinds of aggregation queries over historical stream data. Fourth, we consider carefully about the adaptive factors: runtime memory usage and time latency based on selectivity and processing time, in order to adapt changed factor of stream data and running environments. Then we propose an adaptive query processing strategy. Finally, we implement the data stream management system prototype named SCOD, introduce the system structure and the function of main modules, and give out the main data structures and algorithms.