hive - HDInsight scalability when using Azure Storage

Question

Hi I'm playing around with HDInsight. I'm putting log files into Azure storage and then using Hive external tables to map onto them. I believe Microsoft recommend Azure storage to HDFS so you can delete and recreate the clusters without losing data. What is the scalability vs HDFS. My understanding of HDFS is that it is spread over multiple nodes to allow parallel processesing how does this compare to Azure storage.

Answer 1

在 HDInsight 上，HDFS 存储基于在 VM 的物理主机中运行的磁盘（PaaS VM 在 Windows Azure 中称为辅助角色）。

Windows Azure 存储有自己的可扩展性机制。此处记录了可伸缩性目标：http: //msdn.microsoft.com/en-us/library/windowsazure/dn249410.aspx 让您了解一下，Windows Azure 存储是 Windows Azure IaaS VM 的 OS 磁盘所在的位置。