MapReduce: A key concept of Big Data and Analytics simply explained

 MapReduce in Big Data and Analytics

At times, we come across a number of Technology related terms. We find material on it on the Net, Google etc.
Yet we find out that all the stuff is too technical and not suitable for explaining to a person not in related field in a simple way.
One such concept is that of MapReduce in Big Data and Analytics.

Example Of MapReduce in Big Data and Analytics.

Here I try to explain it in simple terms so that even a non-technical guy can get a brief idea of what it is :

1)As an example : The Input Data(usually Millions/Billions or even more records, or data values) are “Mapped” to a Key/Value Data sets.
For e.g- If we have an input of a site and number of visitors time interval wise:

Time-Interval| No. of Visitors
10:00 -12:00 | 2700
12:00 -14:00 | 1155
14:30 -17:00 | 800
17:00 -19:00 | 5000
19:00 -21:00 | 7550
21:00 -23:00 | 6100

This data will be mapped to Key/Value sets :

{10:00-12:00,2700}, {12:00-14:00,1155}, {14:30-17:00,800},
{17:00-19:00, 5000},{19:00-21:00,7550},{21:00-23:00,6100}

2) Next, Multiple “Reduce” jobs, most probably in parallel shall be started with the aim of further breaking the above input(Millions or more such inputs in real life scenario) and process them.
Finally the results of all such “reduce” jobs are combined to get the results.

The above are accomplished on cost-effective “commodity” servers and many Organizations are increasingly using these systems to do complex time consuming data operations on TeraBytes or even Petabytes bigger data, within acceptable time range, and thus creating value and generating revenue, and saving costs.

Source partially referred :
My article:
http://www.wikinut.com/explaining-the-big-data-and-analytics-concept-of-mapreduce/2k9dqhjw/v-nu7ugr/
on Wikinut