Multiple Continuous Queries Evaluation over Data Streams

Query processing for data streams should be continuous and rapid, which requires strict time constraint. In most previous researches, in order to guarantee this constraint, the evaluation order of join predicates in a continuous query is optimized by a greedy. However, the greedy strategy traces only the first promising plan, so that it often finds a sub-optimal plan. This paper proposes an improved scheme called an Adaptively Sharing-based Extended Greedy Algorithm(A-SEGO). Given continuous queries with multiple join operations, they simultaneously trace a set of promising plans to reduce the possibility of producing a sub-optimal plan. Also it can control the time to optimize continuous queries depending the current processing load by controlling the number of traced plans. Experiment results illustrate the performance of the A-SEGO in various stream environments.