| The rapid developments of computer and Internet technologies have promoted great changes on current network environment and applications. As a result, there are some limitations in the system design current computer networks, which cannot be solved fundamentally just through supplementary to the existing network. The goal of " Universal trusted network and universal service system basic research" is to provide diversified network and terminal access on the universal network platform, ensuring security and mobility of information interaction and provide universal service. In order to realize the intelligence of universal network, the current information of networks need to be collected and analyzed. Intelligent decisions are adopted to realize the network self-maintenance and self-optimization according to these information and changes of network environment. Users are the behavior producers and real-time function main body, so it is important to define a complete model to represent user behaviors. As an important part of the universal network, user information and user behavior information are collected by an effective strategy and analyzed with an efficient analysis model to acquire user interests and behavior characteristics, to further support the intelligent decision of universal network.Based on the investigation of those existing relatively mature user models, user information management systems and behavior analysis technologies on computer networks, an integrated network-based user management system is proposed according to universal network users and network characteristics. The system is divided into five parts:User Registration module, User Authentication module, Behavior Analysis module, Label generation module and Interface Display module; they realize the system function by modularization design, complete the system function and performance test and display certain user information through Web pages. According to the function demands for each module, the following research results are achieved in this paper:1. User Access Authentication module:Both user registration and user authentication modules are belonging to this access authentication module. Considering usability, maneuverability and scalability demands of the system, user access authentication function is based Web push technology.2. User Behavior Analysis module:This module is divided into data acquisition part and data analysis part. Data acquisition module is deployed on access routers, gathering data rapidly using PFRING and deriving application layer information into a database using IPFIX. Data analysis module introduces Hadoop processing architecture, together with related components HDFS, HBase and MapReduce.3. Label Generation module:This module takes user basic information database and user behavior information database as data sources and then extracts its needed information to generate descriptive user tags and show in string form. On the other hand, label generation module completes behavior information receiving, behavior information update and returns the updated results.4. Interface Display module:This module displays user database information through the Web interfaces. Besides, it can show TopN data of flow and applications in statistics. |
