Bibhudutta Jena

Bibhudutta Jena

Bibhudutta Jena

Enhance The Performance Of Name Node By Aggregator Assisted HADOOP Framework

Due to the increased market competition increased data management and analysis has landed as in an era that requires further optimization data management and analysis. Big data means a large amount of generated and used data. This large volume of data stored as a collection of large datasets were not able to be operated using basic computing techniques. It is not simply a data, but it has become a veritable research area and a full fledge subject, which consists of various tools, techniques and frameworks. It embroils the data fabricated by different devices and applications Big data technologies like apache HADOOP provide a frame work for parallel data processing and generation of analyzed results.MAPREDUCE method is used for analysis of data using various data analysis algorithms like clustering, fragmentation and aggregation. A single node, termed as NameNode of HDFS keeps the metadata of the entire file system and monitors the file content placement policy of the data storage subsystems, called DataNodes. As per the existing HADOOP architecture the data received from client is distributed to various data node by the name node and it is the responsibility of name node to track the task being performed by a data-node through a task-tracker. Hence the burden in Name node demon is higher in comparison to the other demons present in the Hadoop framework.

The presented proposal aims to reduce the burden on name node in the HADOOP architecture by providing the assistance through aggregator node which, act as an interface between the name node & data node. In the proposed framework if the load is more on Name node then the aggregator nodes take the responsibility for assigning the data among data node and it also responsible for taking the reduced output from the data nodes and provide the reduced output to the user. Therefore it is helpful for overcome from the single point name node failure which is a major issue in existing Hadoop framework. In the presented paper the framework has implemented in HADOOP framework and it also provides a comparison between the existing Hadoop framework and proposed Hadoop framework on the basis of execution time by taking an predefined data set. Finally the paper has concluded that Efficiency of the proposed work can be improved by adding more number of aggregate nodes in between name node and data nodes.