Limitations and challenges of HDFS and MapReduce

Over these past 6 years, Hadoop has become a highly popular solution to store and process a large amount of data for analysis purpose. Those 6 years of utilization along with the researches undergone which focused on Hadoop enable researches to have a good overview of its advantages, drawbacks and limitations in order to improve the solution initiated introduced. Even though Hadoop 2.0 released in 2012 brought several improvements, especially regarding Hadoop Distributed File System (HDFS) availability and the cluster resource management during MapReduce job execution through the YARN architecture, numerous scopes of improvements are yet to be explored. This paper aims to present those drawbacks and limitations of Hadoop 1.0, explain what brings Hadoop 2.0 and what these remaining scopes of improvements for Hadoop are.