|
With the development of network, communications equipment popularity , a new data-intensive applications gradually emerges, which mainly include: financial analysis, network monitoring , communications, data management, sensor network data processing. In these applications, the data is multidimensional , continuous, rapid , time-varying , while these applications in addition to the traditional ad hoc queries , more likely to require continuous queries . Traditional database technology , because of its assumptions on the data and the corresponding query processing techniques , has been unable to adapt to the requirements of these applications , so the development of a new data management system to effectively meet the requirements of this new type of data processing has become the database research one of the hotspots . The purpose of this paper is to study the data stream management system to be implemented and , in doing major work : study of the data stream data model and query language ; designed a data stream management system prototype GStream; studied ducted filter order problems proposed adaptive algorithm based on greedy thoughts AFO; pair of k- constrained classification , examine the various k- constrained computing role in the connection , put forward the corresponding applications and monitoring algorithm ; MJoin method using multiple connections to achieve operational introduced the concept of intermediate results , an improved multi-channel connection to technology ; finally realized GStream, and the experimental analysis of a variety of algorithms .
|