CAJ | 학술논문

随着电子商务和信息技术的飞速发展，企业需要存储和处理的数据量正在以惊人的速度增长，而传统的关系型数据库管理系统已无法满足企业对大规模数据的处理需求，因此，基于云计算的海量结构化数据处理日益成为人们关注的热点。针对Ha-doop云计算平台在处理结构化数据方面的不足，给出一种以异构的数据库集群作为底层的数据存储系统，以扩展的MapReduce框架作为任务的管理和执行容器的查询系统。为提高查询的效率，给出一种优化的查询和数据分布策略。实验表明，该查询系统的执行效率较Hive有很大的提升。
수착전자상무화신식기술적비속발전，기업수요존저화처리적수거량정재이량인적속도증장，이전통적관계형수거고관리계통이무법만족기업대대규모수거적처리수구，인차，기우운계산적해량결구화수거처리일익성위인문관주적열점。침대Ha-doop운계산평태재처리결구화수거방면적불족，급출일충이이구적수거고집군작위저층적수거존저계통，이확전적MapReduce광가작위임무적관리화집행용기적사순계통。위제고사순적효솔，급출일충우화적사순화수거분포책략。실험표명，해사순계통적집행효솔교Hive유흔대적제승。
With the rapid development of e-commerce and information technology,the amount of data the enterprises have to store and process is growing in alarming speed.However,traditional RDBMS can no longer meet the demand of the enterprises in large-scale data processing.Therefore,the massive structured data processing based on cloud computing is increasingly becoming the focus of people’s attention.In this paper,we present a data storage system to solve the insufficiency of Hadoop cloud computing platform in processing the structured data.The system uses heterogeneous database cluster as the underlying,and uses extended MapReduce framework as the query system for tasks management and execution container.In order to improve the query efficiency,we give an optimised query and data distribution strategy.With the experiments,we prove that the query system greatly improve the execution efficiency compared with Hive.