Survey on Map Reduce Scheduling Algorithms in Hadoop Heterogeneous Environments
R. Nirmalan, K. Gokulakrishnan · 2018
over past few years there is continuous increase in computational capacity and hence there is enormous data flow (i.e., Big Data) which surpasses the capacity of traditional processing tools. These processing demands of Big Data in real time applications produce a big challenge to attain required levels of performance. Map reduce is one of the most efficient parallel distributed programming models for maintaining huge unorganized data sets present in cloud applications. This Map reduce model is implemented in Hadoop which is an open source Java based programming framework. It is widely used for large and high performance data processing with low response time in Big Data. The Hadoop implementation is made on Homogeneous Environment. Since, it reduces the overhead in data transmission and each participating cluster node possesses ideal computing speed and workload as well. However in real time applications the processing nodes may have unique computational capacity and workload emerging in a heterogeneous environment. The regular Hadoop implementation does not produce the required performance into heterogeneous environment. This paper presents a survey on various Map reduce scheduling algorithms with its taxonomy, advantages and disadvantages in heterogeneous environment.