Glister: A framework for iterative MapReduce

In cloud computing era, Big Data analytics is used predominantly. To analyze Big Data we need powerful tools and distributed computation. Map Reduce is one of the programming models that can be adopted by any programming language and used to develop software for analyzing such a huge amount of data. Map Reduce is generally used for parallel processing. But if it can be extended to implement iterative algorithms, performance can be increased. There is a need of some programming model to analyze Big Data that can work on iterative algorithms that may require parallel computation also. For iterative algorithms, some parts that can be parallelized can be implemented in Map Reduce. In this paper, a new framework for Map Reduce is proposed and implemented that supports iterative computation efficiently.