Characteristic Analysis Of Mobile Broadband Data Traffic On Cellular Networks

Internet backbone traffic mixes different client traffic from fixed broadband access and 3G/4G cellular network access.It is difficult to distinguish the two correctly using traditional traffic characteristics and analysis without relying on application layer information and viewing datagram content.Aiming at the requirement of traffic classification based on network access type,this paper proposes a method to analyze the characteristics of mobile broadband data traffic in cellular networks,and constructs a classification model to tell apart the data generated in 3G/4G and fixed broadband access environment flow.This paper mainly analyzes the influence of the delay technology on the IP datagrams of the cellular network,and constructs six delay-related network traffic characteristics in combination with the passive measurement method of TCP/IP protocol and delay in 3G/4G cellular and fixed broadband access traffic sources.On this basis,a supervised machine learning method is used to establish the classification model of network traffic,and to cross-validate the model with public traffic data.The experimental results show that the flow description model established based on these delay characteristics can effectively discriminate the data traffic from the cellular network access or fixed broadband access,the correct rate is more than 92%,and has good coverage and fault tolerance.The main research work of this paper includes the following points:(1)Based on the impact of network architecture on the delay,this paper analyzes the similarity and difference between the cellular network,WLAN and fixed broadband,and studies its influence on the delay characteristics in the network data traffic,and constructs the traffic characteristics of the cellular network.(2)According to the characteristics of cellular network,WLAN and fixed broadband network access traffic,the corresponding statistical methods and criteria of feature statistics are put forward,and the process and method of feature extraction in the mixed traffic flows are summarized.(3)Using the supervised learning algorithm to construct the traffic classification model,and with the the public data set to carry out the analysis experiment,the effectiveness of the extraction feature to the traffic classification is verified,and the advantages and disadvantages of the the advantages and disadvantages of various traffic characterization and classification models under different application conditions are analyzed and benchmarked.