WIN: an efficient data placement strategy for parallel XML databases

The basic idea behind parallel database systems is to perform operations in parallel to reduce the response time and improve the system throughput. Data placement is a key factor on the overall performance of parallel systems. XML is semistructured data, traditional data placement strategies cannot serve it well. In this paper, we present the concept of intermediary node I Node, and propose a novel workload-aware data placement WIN to effectively decluster XML data, to obtain high intra query parallelism. The extensive experiments show that the speedup and scale up performance of WIN outperforms the previous strategies.