基于Hadoop的大数据查询系统简述

近年来，随着计算机技术的迅猛发展，其领域迎来了大数据时代。随着大数据的出现，传统的关系型数据库已经不能满足高储存量的要求，此时成本低廉、有着良好并行性和伸缩性的云数据库应运而生，它采用键值对数据模型和分布式的计算环境。但是海量数据在Key-value数据库中的查询效率低下、实时性差等问题又普遍存在。为了解决查询效率低下这一问题，将多维数据模型和索引技术应用于Key-value数据库，将事实数据以多维的形式进行存储并在多维模型上建立索引以加快查询速度。论文将系统地描述多维数据模型的建立和索引技术的实现，最后简单地和主流Key-value数据库进行优缺点对比。

A Brief Introduction Hadoop-based Big Data Query System

In recent years, with the rapid development of computer technology, its area ushers in the era of big data. With the emer- gence of big data, the traditional relational database connot meet the needs of high storage capacity when the low cost cloud database with good parallelism and scalability comes into being at the historic moment. It uses a Key-value data model and the distributed computing envi- ronment. But the problem that query of huge amounts of data in the Key-value database lacks efficiency and has bad real-time performance is universal. To solve low efficiency in query, this system applies multidimensional data model and the indexing technology to the Key-value da- tabase, storing fact data in forms of multidimensional data and indexing on the multidimensional model in order to speed up the query. The establishment of multidimensional data model and the implementation of indexing technology is systematically described. Finally the thesis briefly compares it to the mainstream Key-value database in terms of advantages and disadvantages.