Data Lake for Enterprises

HDFS

As detailed earlier, Hadoop follows the Master-Slave architecture pattern for both data storage and computing. For data storage, it uses HDFS as the main component. Two sub-components, namely NameNode and DataNode, are present in master and slave nodes respectively. The DataNode stores the application data and NameNode stores the filesystem metadata. The communication between NameNode and DataNode is through TCP-based protocols and is quite reliable and high-performant.

The following figure shows the Master-Slave architecture used in HDFS, with the NameNode and DataNode components:

Figure 09: Master-Slave architecture of HDFS showing ...