| 摘要: |
| SOH（SQL over HDFS）系统通常将数据存储于分布式文件系统HDFS（Hadoop distributed file system）中，采用Map/Reduce或分布式查询引擎来处理查询任务.得益于HDFS以及Map/Reduce的容错能力和可扩展性，SOH系统可以很好地应对数据规模的飞速增长，完成分析型查询处理.然而，在处理选择型查询或交互式查询时，这类系统暴露出了性能上的缺陷.提出一种通用的索引技术，可以应用于SOH系统中，以提高其查询处理的效率.分析了SOH系统访问HDFS文件的过程，指出了其中影响数据加载时间的关键因素.提出了split层和split内部双层索引机制；设计并实现了聚集索引和非聚集索引；最后，在标准数据集上进行了大量实验，并与现有基于HDFS的索引技术进行了比较.实验结果表明，所提出的索引技术可以有效地提高查询处理的效率. |
| 关键词: 查询 索引 HDFS SOH 大数据 |
| DOI：10.13328/j.cnki.jos.005161 |
| 分类号:TP311 |
| 基金项目:国家重点研发计划（2016YFB1000702）；中国人民大学预研委托（团队）基金（14XNLQ06）；国家自然科学基金（61003086） |
|
| Multi-Layered Index for HDFS-Based Systems |
|
HE Long1, CHEN Jin-Chuan1,2, DU Xiao-Yong1,2
|
|
1.School of Information, Renmin University of China, Beijing 100872, China;2.Key Laboratory of Data Engineering and Knowledge Engineering, MOE(Renmin University of China), Beijing 100872, China
|
| Abstract: |
| The SOH (SQL over HDFS) systems usually store the data into distributed file system HDFS (Hadoop distributed file system), and process queries by the Map/Reduce computing framework or distributed database query engine. Benefitting from the fault tolerance and scalability provided by Map/Reduce and HDFS, SOH systems perform well in processing analytical queries over big data. However, the efficiency of such systems is too low to meet the requirement of selective queries or interactive queries which have strict limit on the query response time. This paper proposes a HDFS-based index, called HIndex, for SOH systems. HIndex can easily be integrated into the existing SOH systems to improve the efficiency of query evaluation. The process that SOH systems access data stored in HDFS is analyzed, and the important factors affecting the time cost is highlighted, a two-layer index structure is proposed, and both aggregated and non-aggregated index techniques are implemented. According to the experiments conducted on standard datasets, HIndex performs much better than Hadoop++, a state-of-the-art HDFS-based index. |
| Key words: query index HDFS SOH big data |
