Network traffic analysis based on Hadoop

With the rapid development of mobile Internet, numerous mobile services have become part of the people´s life. Accordingly, the traffic rate generated by mobile devices are humongous and it is very hard to analyze its characteristics in traditional ways due to their scalability. This problem can be viewed as a big data problem and we develop a Hadoop-based system, Traffic Analysis System (TAS), to as the solution in this paper. Based on TAS, we analyze the real mobile Internet traffic data collected from a capital city in southern China. The analysis results illustrate the basic makeup of current mobile devices. In addition, it indicates that the users with different interests in different networks have quite different behaviors, which could provide important references for Internet service providers.