一种并行XML数据分片算法的设计与实现  

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

作　　者：汤南[1] 于亚新[1] 王国仁[1] 于戈[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110004

出　　处：《小型微型计算机系统》2004年第7期1164-1169,共6页Journal of Chinese Computer Systems

基　　金：教育部高等学校优秀青年教师教学科研奖励计划基金资助项目;教育部高等学校骨干教师资助计划资助项目;国家自然科学基金(60 173 0 5 1;60 2 73 0 79)资助

摘　　要：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 .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.

关 键 词：并行数据库 XML文档 分片 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]