Analysis of network node behavior based on MapReduce

With the widespread popularity and rapid development of Internet, the network is becoming more complex. Analysis of network node behavior can be crucial for network management and service provisioning. In this paper, we conduct the detailed measurement analysis of the network node (NN) traffic characteristics based on a large-scale data set collected from an enterprise network traffic monitoring device, which is deployed at the entry of enterprise network covering the Guangdong province in China. We analyze the distribution of NN traffic volume and visitor number. We discuss the rank of NN in terms of traffic volume distribution and visitor number distribution, which follows power law distribution. Furthermore, we illustrate the traffic characteristics of NN in terms of visitor number. To handle the problem of big data processing, we make use of Hadoop framework and MapReduce programming model to analyze huge traffic data, which handles a large collection data efficiently and quickly for scalable storage and analysis. Thus, in this paper, we propose a network traffic analysis method based on MapReduce in enterprise internet.