HDFS typically stands for Hadoop distributed file system.As compared to traditional RDBMS, HDFS follows the distributed approach and has the following advantages:

1)  Data doesn't need to be centralized. However,it is distributed across network.As per traditional RDBMS, data needs to be first accumulated at one place with the help of relational tables.In HDFS, there is no need to accumulate data at one place.Instead,job is split and executed parallely at different nodes where data is stored.

2) Fault tolerant: In HDFS, data gets replicated depending on the replication factor.By deault, it is 3.

3) HDFS works well with structured as well as unstructured data while RDBMS is designed only for structured data.

4) RDMS is by and large utilized for OLTP processing while Hadoop is right now utilized for logical and particularly for BIG DATA handling.

Example:
Plotting a month to month electricity use of a client by looking at between earlier months, between his or her neighbors or even between clients on the same lanes. This will bring more mindfulness, however running such complex correlation by dissecting vast arrangement of information takes a few hours of preparing time, and introduction of Hadoop help enhancing the processing execution from 10 times to 100 times or more.