HBase High Performance Cookbook

Using the filesystem

HBase depends on the Hadoop Distributed File System (HDFS).

HDFS fundamentally is a distributed file system, which relies on following core principles:

Getting ready

The following are the benefits of using HDFS:

• It's designed to work as a fault-tolerant system and is rack aware.
• It works on the low-cost commodity hardware.
• HDFS relaxes core system POSIX requirements to facilitate streaming access to the underlying OS access of file system data.
• It's designed to write once and read many times. It also supports parallel reading and processing the data (read, write, and append). It doesn't support random writes of data.
• It's designed to scale at a very large level, which means file size like petabyte of data.
• It works with minimum data ...