Recovering Nodes Failure in Replica System by Using Data Replacement Algorithm in a HDFS Cluster

Abstract

HDFS (Hadoop Distributed File System) is designed to store big data reliably, and it has been distributed file. However, it is not replacement for failure nodes. Replication technology is a key component in every computing enterprise as it is essential to applications such as backup, file distribution, file synchronization as well as file sharing collaboration. The replacement concept is simple but its algorithmic optimization and system implementations are challenging and difficult. Data Replacement Algorithm was implemented to make the replicas to get placed on failure nodes. As this system can reduce data failure and cannot affect other Data Nodes, it is more reliable and usable.