A Cloud Platform for flow-based analysis of large-scale network traffic

Traditional flow-based network monitoring and analyzing tools are challenged with large-scale network traffic for the big size of data. This paper proposes a Cloud Platform for large-scale network analysis. It exploits HDFS and DataX to deal with the storage of heterogeneous big data. And the MapReduce model is applied to concurrently compute the analyzing and correlating on large amount of network flow data. Flow-based techniques are employed in collecting network data to make possible an application level analysis of network usage. Experimental results show that the performance of the platform is greatly improved compared with that of the traditional RDBMS application.