Please use this identifier to cite or link to this item:
Title: Decoupling Storage and Computation to Run Hadoop on Grid using Ubiquitous Erasure Codes
Authors: Raj, S.
Keywords: Computer Science & Engineering
Issue Date: 2015
Abstract: With the advent to computing Grid more and more complex applications on large data sets is envisaged. MapReduce has become the leading programming model for processing large data sets with parallel, distributed algorithms on a HDFS cluster. In this paper we propose a novel method to integrate Hadoop with grid. To extend the MapReduce beyond cluster to the grid, we propose a model for decoupling storage and computation and use ubiquitous Erasure Codes for improving fault tolerance and reducing redundancy and NameNode metadata. A detailed analysis is done on the effects of the changes proposed on traditional Hadoop framework.
Appears in Collections:01. CSE

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.