Design and Implementation of Network User Behaviors Analysis Based on Hadoop for Big Data

The network user behaviors analysis under the big data environment is attractive to network security recently for that it can discover the abnormal user behaviors to prevent the potential threats. However, the user behaviors are dynamic which is difficult to capture the users' comprehensive behaviors in a single device by capturing or collecting the static dataset. More specially, the increase of the network users, network traffic and network services bring many challenges such as fast data collection, processing and storage. Therefore, we propose and implement a network user behaviors analysis system in this paper, which is based on the Hadoop distribution platform to capture the traffic and analyze the user behaviors in terms of the search keywords, user shopping trends, website posts and replies, and web visited history to acquire the uses' dynamic behaviors. To evaluate our system, we capture the packets in the campus networks, and the results show that our system can capture the users' long-term behaviors and acquire the user behaviors in detail.