题目

英语翻译
Hadoop是云计算平台的实现框架,是一个Apache公司开源软件实现基于java语言的云计算开发平台,也是数据处理软件平台,包括分布式计算的海量数据的大型计算机集群的大规模开发与运营.通过分析云计算平台Hadoop 在面对海量小文件所存在的问题,可知Hadoop中的主节点NameNode把文件及文件夹等的所有元数据放在内存中进行管理,大规模的小文件会使得主节点NameNode的内存利用率大大的降低甚至导致系统崩溃.为了解决这个问题,本文结合云平台Hadoop自带的小文件解决方案,提出了一种基于Sequence File与预取机制相结合的Hadoop海量小文件解决方案.实验表明,在应用该方案的Hadoop云平台中,文件的访问速度提高了约34%,主节点NameNode服务器内存的使用率降低了约15%.

提问时间：2020-11-01

答案

Hadoop is implementation framework of cloud computing platform,is an Apache open source software implementation based on Java language cloud development platform,and the data processing software platform,including the massive amounts of data distributed computing of large mass development and operation of computer cluster.By analyzing the Hadoop cloud computing platform in the face of a huge number of the problems existing in the small file shows the Hadoop master node NameNode put files and folders,etc all metadata manage memory,large-scale small file will make the NameNode winner node memory utilization greatly reduce or even lead to system crash.In order to solve this problem,based on the cloud platform Hadoop's own small File solution,is put forward based on the Sequence File and prefetching mechanism of combining the Hadoop mass solution for small files.Experimental results show that the application of the scheme Hadoop cloud platform,file access speed by about 34%,the master node NameNode server memory usage reduced about 15%.