Date Added: Jan 2011
In this paper, the authors design and evaluate an efficient and dynamic regrouping approach to optimize a large continuous query workload. The key idea of the authors' regrouping algorithm is to find a best solution by removing redundant groups within the existing solution from incremental group optimization. Since the existing query groups usually constitute a very small portion of the entire search space, such a heuristic-based approach is key to achieve an efficient regrouping algorithm. In addition, timing of regrouping is also critical to the efficiency of regrouping. Overall, the authors believe that constant incremental grouping in conjunction with occasional dynamic regrouping can achieve a high-quality grouping at a fairly low cost for optimizing a large, dynamic continuous query workload.