摘要
提出新的小枝模式索引方法和无回溯的标记过滤算法,通过增加较少的系统存储资源提高小枝模式订阅的过滤计算速度。过滤算法对订阅的每个谓词单独设立标记,从而简化匹配运算,通过对标记组的并操作使算法无须回溯即可实现过滤。实验结果证明,该算法性能较好,适用于上百万订阅量的数据流系统。
This paper proposes a new twig pattern index algorithm and a non-retracement tagging filtering method.It enhances filtering computing speed of twig pattern subscribe by increasing a few storage resources of the system.A tag is set for each predicate in the query to simplify matching operation,and through AND operation to filter the queries without retracement.Experimental results show that the algorithm has good performance and adapts to the system with millions of subscriptions.
出处
《计算机工程》
CAS
CSCD
北大核心
2010年第9期50-52,共3页
Computer Engineering
关键词
数据流
发布/订阅
小枝模式
data stream
publish/subscribe
twig pattern