INTERNATIONAL JOURNAL OF ENGINEERING, SCIENCE AND - Volume 3, Issue 4 , December 2014

Pages: 217-228

Date of Publication: 01-Dec-2014

Print Article   Download XML  Download PDF

A MODERN APPROACH ON MAP REDUCE CLUSTER USING DYNAMIC SLOT ALLOCATION OPTIMIZATION FRAMEWORK

Author: B.Kayalvizhi* S.Karpagaselvi*

Category: Engineering, Science and Mathematics

Abstract:

Emerging technology and effective mechanism oriented approach makes Map Reduce a popular as well as important computing paradigm which dealing with big data process in cloud computing. The mechanism of un-optimized resource allocation is the severe drawback for slot based Map Reduce System. To overcome the failures of the optimize resource allocation we identify three aspect in this process in order to improve the resource allocation in an effective manner. The initial one is pre-configuring the map slots by reduce the slots which are under-utilized and fungible. The key behind these methods is it makes the map slots fully utilized which reduces the empty slots and vice-versa.

Keywords: Slot Allocation, Slot Pre-Scheduling, Map Reduce, Delay Scheduler, Hadoop Fair Scheduler, DynamicMR