Design And Implementation Of Continuous Query Processing Prototype System Over Data Streams

Recently there has witnessed a focus on processing new data-intensive applications, such as financial service, network monitoring, telecommunications data management, sensor detection, and so forth. The common feature of these applications is that, the data generated in these applications are more suitable for being modeled by transient data streams than by permanent relations. In the model of data streams, a data unit could be a tuple of related data, such as network measurement, call records, sensor reading, and so on. For the sake of the characteristic of huge size, high speed of data arriving, time-varing, the data in data stream model could not be managed by traditional database.On the basis of studying the new theory and technology of data stream management, a new data stream management system prototype based on preprocessing the data streams by hardware, which named SEUSTREAM, is developed. Firstly the architecture of SEUSTREAM and the preprocessor of data streams are introduced. Then the paper gives the design of a continuous query language X-SQL, which support the continuous query on data streams. And then a compiler of X-SQL is introduced. Its target machine is the data stream preprocessor in SEUSTREAM. In SEUSTREAM, users submmit continuous queries of data streams by X-SQL language. After lexical analysis, sytax parsing, semantic analysis and query optimization, the queries commited by users would be transferred into an abstract syntax tree (AST) by the compiler. And then the compiler would generate machine code of the preprocessor.