Research On The Query Processing Technologies Of Distributed Data Stream

With the recent development in the field of information management and application, such as Internet monitoring, Web application, sensor network, and financial analysis, a new form of data model, the data stream, has come under increasing attention. Data stream features continuous arrival of data, with speed and very large scale. The study on data stream is one of the hot topics among the database circle all over the world recently. Conventional database technologies have been well developed and widely applied. Unfortunately, they could not be adopted to handle data stream.Many applications of data stream are naturally distributed and they are very suitable for distributed processing. Distributed processing is a very promising route towards a more effective and adaptive data stream processing model. This paper focuses the research on the query processing technologies of distributed data stream.Firstly, the distributed data stream query processing model concept and the existing query operator placement of distributed data stream technology are introduced in detail. On the foundation of analyzing the algorithm based on spring relaxation of query operator placement of distributed data stream thoroughly, an improvement to the algorithm through altering the object function optimized by the algorithm and initialization of coordinate. And experimental results show that the improvement is effectively.Secondly, query plan is defined and two greedy algorithms are proposed to produce optimized query plan in distributed data stream management system. In addition, a query reusing algorithm is proposed to reuse the overlapped query in distributed data stream management system.Finally, on the basis of analyzing the exited typical data streammanagement systems, a generic distributed data stream management system, named GDDSMS, is developed.