2

Big data has been an imperative quantum globally. Gargantuan data types starting from terabytes to petabytes are used incessantly. But, to cache these database competencies is an arduous task. Although, conventional database mechanisms were integral elements for reservoir of intricate and immeasurable datasets, however, it is through the approach of NoSQL that is able to accumulate the prodigious information in a proficient style. Furthermore, the Hadoop framework is used which has numerous components. One of its foremost constituent is the MapReduce. The MapReduce is the programming quintessential on which mining of purposive knowledge is extracted. In this paper, the postulates of big data are discussed. Moreover, the Hadoop architecture is shown as a master- slave procedure to distribute the jobs evenly in a parallel style. The MapReduce has been epitomized with the help of an algorithm. It represents WordCount as the criterion for mapping and reducing the datasets