Reducer

After Shuffle, the intermediate output will then be passed on to Reducer as an input. Reducer mainly performs an aggregation function on the key-value pairs for various calculations depending on the program's purpose. The aggregation can be of any type, including summation, multiplication, and so on, as illustrated in the following figure:

Finally, the program output is then written to the HDFS file system. It is worth noting that the number of reducers is configurable in the Hadoop system; by default there is one.