The above figure IV shows the Hadoop architecture
in a cluster setup which has three nodes. Form the
above figure we can see JobTracker is
communicating with the TaskTracker for assigning
jobs such a Map, Reduce and Shuffling. NameNode
is responsible for handling all the data nodes in a
cluster and DataNode is assigned to each node in the
cluster. In addition to most important NameNode
there may be a less important NameNode in the
Master node which is typically used to store the
replica of metadata of the NameNode for handing
Letdowns