Students Online Behavior Analysis System Based On IPv6 Campus Network

With the rapid development of IPv6 network technology,the network application is more and more extensive,students ' learning and entertainment cannot be separated from the support of campus IPv6 network,and also make the campus network export data flow more and more.In order to understand the health of the students' surfing habits in the campus network,and to ensure the accurate configuration of the campus network resources,it is necessary to analyze the usage of the network traffic accurately,that is to do the content monitoring and auditing of the campus network.The prerequisite of network monitoring and audit realization is to capture and analyze the data of campus network export.At present,the high speed and accurate data packet analysis and processing system is becoming more and more important to the campus network resources and the students ' Internet behavior management.The campus network data analysis processing system mainly includes the campus network data capture,the solution package and the application system classification and so on function.Based on the demand analysis,this system finishes researches below.(1)Research on the capture of fast packet based on DPDK.Traditional algorithm on the capture of fast packet will lose packet when it tries to capture the campus network traffic which is above 1Gpbs.This paper conducts research on the algorithm of the capture of fast packet based on DPDK through DPDK mechanism,and eliminates unnecessary system calls and multiple copy among different regions of memory which significantly improves the efficiency of the data packet capture.This algorithm has been proved by experiments that it can capture the data packet of the ted thousand mainstream campus network.(2)Research on the parsing algorithm of data packet based on DFI and DPI.Due to the relatively high complexity of DPI and the low accuracy rate of identifying p2 p flow and traffic camouflage,this paper restored the stream based on IPv6 flow label,classifies data stream by method based on DEI of SVM,especially the web traffic and other traffic,then extracts domain names of web traffic through DPI and next conducts the application recognition and classification based on WWW.This algorithm has been proved by experiments that it improves the speed and accuracy rate of application recognition.(3)The realization of client of online behavior presentation.This paper fulfills the visual presentation of analysis results by developing a student online behavior analytic system based on JAVA Web.Through this system,the student affair officials can understand the online behavior of students,and thus supervise the students to form healthy online habits.The network managers can also learn more about the use of campus network,which provides data support for reasonable allocation of network resources.