Downloads: 109 | Views: 155

Survey Paper | Computer Science & Engineering | India | Volume 5 Issue 6, June 2016

Performance Analysis for Optimizing Hadoop MapReduce Execution

Samiksha Misal | P. S Desai

Abstract: The Apache Hadoop data changing writing computer programs is doused in an intricate situation made out of gigantic machine bunches, limitless data sets, and a couple taking care of vocations. Managing a Hadoop situation is time escalated, toilsome and obliges expert customers. Likewise, nonappearance of learning may include misconfigurations adulterating the gathering execution. To address misconfiguration issues we propose an answer completed on top of Hadoop. The goal is showing a tuning toward oneself segment for Hadoop businesses on Big Data circumstances. Late years have witness the improvement of distributed computing and the huge information period, which raises difficulties to conventional choice tree calculations. In the first place, as the measure of dataset turns out to be to a great degree huge, the procedure of building a choice tree can be very tedious. Second, on the grounds that the information can't fit in memory any all the more, some calculation must be moved to the outer stockpiling and hence builds the I/O cost. To this end, we propose to execute a normal choice tree calculation, C4.5, utilizing MapReduce programming model.

Keywords: MapReduce, Hadoop, Self-tuning, Optimization, Decision tree

Edition: Volume 5 Issue 6, June 2016,

Pages: 2219 - 2223