Analysing Cascading over MapReduce

Published in Research Journal of Computer and Information Technology Sciences (RJCITS), 2016

Recommended citation: Kaustuv Kunal. Analysing Cascading over MapReduce, proceedings of Research Journal of Computer and Information Technology Sciences (RJCITS), Vol. 4(9), 1-4, September (2016)

[PDF]

Abstract

In recent years Big Data has grown significantly. Hadoop with MapReduce performs distributed processing of large data sets in fault tolerant and cost effective manner. Cascading is an abstraction layer upon MapReduce and allows developers to solve problems using tuples and fields. Many business problem can be solved conveniently with tuple rather than MapReduce key value-pair. The paper advocates cascading over MapReduce and illustrates how lengthy tasks in MapReduce are easily done in Cascading supported by a cloud based architecture and a case study.