The Research Of Data Stream Query Processing Algorithm

Recently, data streams emerge in many applications broadly, such as sensor network, stock analysis and network traffic monitoring . Far different from traditional data, data streams are unbounded, real-time and rapid arrival rate. Data stream query processing has been an active research aspect.In this pape, we conduct data stream query operators algorithms, design and realize several operator algorithms, including operators over non window, such as select, project and union, and operators over sliding window, such as join and intersect.Queries on data streams consist mainly of continuous queries. In. data stream systems continuous queries have two types of execution manner: immediately execution manner and periodically execution manner. This paper presents the barrel sliding window join algorithm, which is based on the improve of general periodically execution sliding window join algorithm and using the idea of deviding the sliding window into barrels. The sliding window will be divided to several barrels and is maintained by barrel, the size of which is equal to the hop. The sliding window join algorithm based on sub-barrels over hop is a sort of symmetric operation algorithm based on pipeline. We present different join algorithm by considering an effect based on the structure of barrel.Theory analyses and experiment results show that the sliding window join algorithm based sub-barrels over hop and the sliding window intersect algorithm based sub-barrels over hop reduce the cost of maitaining the window, consequently , have a better complexity of time.