MapReduce: Architecture and Internals

MapReduce is programming model and implementation for generating and processing large data sets. Users specify map function that processes key function to generate key pair and reduce function that merges all the resulted value with the sane intermediate value. Programs written in this function style are automatically executed parallelly on the large cluster. The run-time system take care of the complexity in partitioning the data input, scheduling the program-s execution across the machine cluster and inter-machine communication. This allows programmer to easily utilize the resources of the whole distributed system. In this paper, we describe the architecture of the mapReduce and its working. We also describe the components of the MapReduce and give its brief description. We have discussed the internals of the MapReduce and explain the working of MapReduce with an example.