一种并行XML数据分片算法的设计与实现

随着XML在各个领域的广泛应用，XML数据规模急剧增加，查询处理也日趋复杂。集中式环境由于单机存在的I／O瓶颈问题，在处理日益复杂的查询操作时已渐渐不能满足应用需求。因此，高性能的并行数据库在XML处理方面的研究，必将为XML数据库的。存储和查询处理带来新的曙光。本文讨论了一种对XML文档的基于节点轮循法进行分片的设计原理及实现算法。实验结果证明该方法具有很好的加速比和缩放比性能。

Design and Implementation of a Parallel Data Partitioning Algorithm for XML Data

With the wide use of XML in many applications over the Web, the scale and the size of XML documents are increasing rapidly and the query processing becomes more complicated than in traditional databases. Centralized environments cannot meet the requirements of Web applications well due to the problem of I/O bottleneck caused by XML documents with large scale and huge size. Parallel query processing is one of promising approaches to solving the bottleneck and data partitioning is one of key issues of the parallel query processing. In this paper we propose a Node based Round Robin data partitioning, short for NRR, to partition a huge size XML document so that queries on the document can be processed in parallel. Our experimental results show that the method has good performance of both speedup and scaleup.