Introduction

As the usage of HBase is becoming a more real-time-based, it's important to make sure that HBase performs the read and write at a consistent speed agnostic to the scale at which it is running:

Type of work	Usage patterns	Remarks
Data intensive	IO, Network	Copying data from HDFS
Text searching	CPU, IO, Network	Locating different sets of data
Machine learning	CPU, IO, Network	Grouping, categorizing, applying models
MapReduce, indexing, grouping	IO, CPU, Networking	Filtering, realigning, combining
Master, slave, region server, Zookeeper communication	Network, CPU, IO	Communicating between different systems, load balancing

In doing so, we need to make sure that all the distributed components perform at a benchmark that is executable on the Linux-based OS.

The list of distributed components is as follows:

• Hardware

• OS (Linux, REH6.5)

• JVM/Java 1.7.0_10 and above

• HDFS

• Read/write on the OS

• Region servers

• Master servers/HMaster

• Client servers/HClient

• Zookeeper

• Table design

• Large...