A New Replica Placement Policy For Hadoop Distributed File System

Today, Hadoop Distributed File System (HDFS) is widely used to provide scalable and fault-tolerant storage of large volumes of data. One of the key issues that affect the performance of HDFS is the placement of data replicas. Although the current HDFS replica placement policy can achieve both fault tolerance and read/write efficiency, the policy cannot evenly distribute replicas across cluster nodes, and has to rely on load balancing utility to balance replica distributions. In this paper, we present a new replica placement policy for HDFS, which can generate replica distributions that are not only perfectly even but also meet all HDFS replica placement requirements