作者: Abhinandan Das , Johannes Gehrke , Mirek Riedewald
关键词:
摘要: We consider the problem of approximating sliding window joins over data streams in a stream processing system with limited resources. In our model, we deal resource constraints by shedding load form dropping tuples from streams. first discuss alternate architectural models for join processing, and survey suitable measures quality an approximation set-valued query result. then number generated result as measure, give optimal offline fast online algorithms it. thorough experimental study synthetic real show efficacy solutions. For applications demand exact results introduce new Archive-metric which captures amount work needed to complete case are archived later processing.