Hadoopdistributed Architecture’ Sresearchandpractical Application

As we all know, Big Data is a very popular IT expression in today’s information technology disciplines. In order to reducethe complexity ofhandlinglarge amounts of data,Apache develops Hadoop--A reliable, scalable, distributed computing framework. Hadoopdistributed architecturecan easily deal withthe challenges ofbig data,which Contains both structured and unstructured data,in twenty-first century.Moreover,Hadoop can process massive data parallely on large-scale too.Thus,it is very important to have a basic knowledge of Hadoop^Hadoop Distributed File System and MapReduce.On the basis of having knowledge of Hadoop Distributed File System and MapReduce,we do research on how to improve Hadoop’s stability,and attempt to optimize its performance.In this thesis,we start from focus on the research background,research content and research purpose,we also make a thorough inquiry into the Indispensable major role of Hadoop in twenty-first century;on the theoretical foundation, Hadoop distributed architecture system is set up,and modules such a HBase,Hive are loaded into the Hadoop distributed architecture system.During this procedure,we also analyze and solve the problems in the building and running process of the Hadoop distributed architecture system.Based on this,we carry on further analyze and research on HDFS,while we also research on the common way to read and write.Finally,we investigate the optimization point that HDFS’s practical application may contain and carry on theoretical study.What’s more,we carry on performance testbased on the previous sduty and the Hadoop distributed architecture system.