Aggregation Query Research Over Continuous Data Streams

With the development of network, telecommunication and sensor technologies, the data is multidimensional, continuous, and quickly changed with time. Data stream theories and technologies have been a new research field of database theory. How to continuously query on the continuous data streams is an important research direction in the data stream theories. In this paper, we focus on the aggregation query on the continuous data streams. First, to deal with the problem of precise query on the data streams, three improved mathematical aggregation query methods which are based on sliding window model are proposed. The three methods can efficiently give the exact query result under the condition of consuming a few memories. Second, the issue of approximate aggregation query on the data streams is discussed. Meanwhile, the approximate query theories based on the histogram synopsis is proposed, and the autosuited histogram synopsis method and the histogram synopsis synchronization method are provided. Third, the problem of predictive aggregate queries on the continuous data streams is researched, and a new predictive aggregate query method based on basic window synopsis and predictive equation is given. According to the theory mentioned above, a new Data Stream Management System Jasper which extends the SQL and forms a new continuous query language JasperCQ is developed. JasperCQ not only supports the continuous query, but also supplies the approximate query and predictive aggregate query. In conclusion, the experimental results show the correctness of the theory.