| With explosive applications of information processing in sensor networks, business processing, network monitoring and security, communication, industry manufacturing, and many other areas, more and more data take the form of multiple, rapid, time-varying, possibly unpredictable and unbounded streaming data (data streams) rather than finite stored data set. In all of the applications cited above, it is not feasible to traditional technology on processing data, so data stream processing technology becomes a new and popular topic in database research area.Now, the research of data stream focus on the general real-time applications. But in some critical task applications (e.g., the monitoring of space craft, online auction), data process must be highly real-time; once missing the deadline, will lead to catastrophic results. The problem of deadline and critical are two major characteristics, so it is necessary to do research for a proper real-time processing strategy in order to improve query efficiency.For some high critical applications, some conceptions (e.g., the deadline, slack and critical) of the real-time system are introduced into DSMS, and give new meanings for them according to the characteristics of continuous query in DSMS. According to these new conceptions, a priority model named CriticalModel for the applications of critical task is proposed. In the CriticalModel, the deadline is nearer or the slack is shorter or the critical is higher, the priority is higher. And a structure of priority tree is proposed in order to realize the unique execution sequence of the query based on priority. A real-time scheduling strategy named RealScheduling based on priority is proposed, in order to increase the Hit Value Ratio (HVR) of continuous query scheduling.The experimental results show that RealScheduling raise the Hit Value Ratio (HVR), and increases the succeed ratio of continuous query scheduling in different environment of data stream. |
PDF Full Text Request