Components of Hadoop Distributed file System (HDFS)

• HDFS is a spread across multiple machines (Simple with commodity Hardware)
• Nothing unique about individual machine but unique part is a cluster as a whole is highly fault tolerant 
• Well Suited for large Batch Jobs
• Not a low latency system
• Data is HDFS is very very large (Semi-Structured)
• Any data in HDFS in split across multiple disks where each disk in present on a diff machine in a cluster
• File system manage machine and space
• Setup by Master-Salve Nodes
• Master Node (Name Node) coordinates with Slave Nodes(Data Node)
• One Namenode/Cluster
• For Example - Name note is like a table of content of a book and data node are the actual chapters
• NameNode has 2 responsibilities 
  • Manage the overall file system
  • Stores (Directory Structure)
  • Other File metadata
• DataNode
  • Physically stores the data