MapReduce is a programming model, also defined as a computer architecture template, designed to perform parallel computations and distributed on very large data. MapReduce is implemented on development platforms called frameworks, the best known to date is Hadoop. But other frameworks also implement MapReduce to take advantage of all the potential performance that enables this architecture model.

MapReduce has an important function in the management of large volumes of data: the distribution of these data on a large number of machines servers. This feature allows parallel processing of multiple tasks. The Map function performs operations dedicated to each element. Reduce the function brings them all together and outputs the result in condensed form.

Here is the diagram of the operation of MapReduce component:

mapreduce hadoop bigdata
MapReduce Architecture

To date, the Hadoop platform is most used in the market, and primarily by the major web players such as Google, Yahoo !, Microsoft, Amazon, eBay, Facebook, Twitter, LinkedIn, Spotify, StumbleUpon …

Hadoop and MapReduce markets are growing rapidly now.

See also: Hadoop Platform