Comparative Analysis of GFS and HDFS: Technology and Architectural landscape

Abstract

Distributed file system has evolved as the De facto file system to store and process Big Data. GFS and HDFS are considered to be the frontrunners and are becoming the favored frameworks options for big data storage and processing. This paper explores the architectural dimensions and support technology of both GFS and HDFS and lists the features comparing the similarities and differences between the two. Architectural exploration is performed with the focus on GFS master, GFS chunkserver and GFS client with respect to the HDFS NameNode, HDFS DataNode and HDFS client. The paper further explorers available support technologies and add-ons for GFS and HDFS.