Static optimization of conjunctive queries with sliding windows over infinite streams

Ahmed M. Ayad, Jeffrey F. Naughton
2004 Proceedings of the 2004 ACM SIGMOD international conference on Management of data - SIGMOD '04  
We study the problem of static optimization of conjunctive queries with sliding window joins over unbounded streaming information sources. While previous work has suggested focusing on maximizing the output rate of queries over streaming information sources, we show that in steady-state, for conjunctive queries with sliding windows over unbounded streams, all feasible plans have the same output rate. For this reason, we suggest that the goal of optimization for such queries should be to
more » ... hould be to minimize resource utilization in steady state. We formulate a cost model for the problem and introduce a framework for optimization based on feasibility and resource management. We then use the model to study where to optimally place random drop boxes and how much to drop in each box in case no feasible plan exists and approximation is necessary. The goal of the approximation is to achieve the maximum throughput of the resulting tuples given the resource constraints. We find that, given a plan to approximate, the optimum placement and values are easily obtained. However, finding the best plan to approximate is not necessarily an obvious task.
doi:10.1145/1007568.1007616 dblp:conf/sigmod/AyadN04 fatcat:6xco5h2vvzbypioiuujmf7xgse