An Adaptive Query Processing Mechanism in Data Stream System
数据流中一种适应性查询处理机制

Aimed at properties of continuous queries, we present an adaptive query processing mechanism. It can not only optimize the time at which the last result tuple appears, but also optimize for the number of answers computed at any specified time after the query evaluation. This mechanism depends mainly on an output-rate-based cost model. Firstly, this model treats changed input rate, predicative selectivity and execution time of operators as its function variables. Secondly, it will compute output rate as function output. So this cost model can continuously adapt to variable factors of environment and data stream itself. Meanwhile, it can be a standard to select query plan. Experiment has proved that this mechanism validly increases throughput and output rate, meanwhile reduces output time latency and memory requirements.