关键词: 数据流, 适应性, 查询计划, 调度

Abstract: Continuous query has some special properties. With this motivation in mind, we think carefully about adaptive factors: run-time memory usage and time latency based on selectivity and processing time. Then we present an adaptive query processing strategy —HoliAdapt. First, we collect continuously statistical information over query windows, then we optimize dynamically query planning according to computing about statistical information. Thereafter, we adopt a scheduling strategy to schedule adaptively operators. Experiments have proved that this query processing strategy can availably minimize run-time memory usage and time latency, and ultimately improve performance of query.

Key words: data stream, adaptability, query plan, scheduling