I have a small doubt regarding reading and writing to HDFS. In general, when the client wants to read/write data from/to HDFS, it will ask the NameNode for block information. Then, the NameNode will give some information about the block that contains the data. Then, the client will directly go to datanodes and do operations.
What if the client doesn't ask the namenode for block information and directly accesses the datanode without any authentication? .
Can someone please explain this?