Design And Implementation Of 4G LTE Network Traffic Analysis System

While the mobile Internet constantly creating new form of industry, the new industries, such as information service, mobile social intercourse, mobile commerce etc. began to develop. Global operators are faced with the huge challenges of the explosive growth of the data flow. In the face of the mass data of mobile Internet, mobile operators, equipment manufacturers, vendors, independent third party application need a set of complete function of traffic analysis system to analyze, inspect and make use of the operational data of the LTE actual network. Such system usually includes the function modules of data acquisition, decoding, storage, statistic and application analysis.According to the functional requirement of 4G LTE traffic analysis system, this thesis mainly discusses the 4G LTE protocol stack decoding and mass flow data analysis, aiming to solve the three issues including traffic decoding, parameter statistics and application analysis.1. The thesis researches the decoding method of TLV/PLC, explores the decoding rule of ASN.1 for special application of LTE network data encoding and decoding, analyzes the hexadecimal code streams. Then transforms them to different ports, protocol signaling and IP data which can be identified by network, terminals and engineering personnel and proposes specific implementation method of PLC decoding rule in LTE network data encoding.2. The thesis recombines the scattered data of user information, application business information, network state information and signaling, proposes control plane relevance on the basis of SIAP internal relevance, NAS bearer relevance and GTP relevance; Then analyzes data of user plane and combines with the AN and CN configuration table to form multi-port relevance scheme of user behavior from control plane to user plane, proposes multi-dimension parameter statistics of user, application, eNodeB and whole network.3. Through using the DPI method such as protocol port identification, keyword matching, flow feature matching, the thesis extracts and analyzes the key information from all kinds of data and to make identification matching of the unknown data as far as possible and helps the business operators to analyze the contribution value of application business, so as to better manage the network traffic on the basis of processing, sorting and analyzing application performance of the identified matching data.The test results show that the 4G LTE traffic analysis system discussed in the thesis can correctly identify the collected data. The statistical parameters can effectively reflect the state of the network. DPI can accurately identify more than 98% packets. The decoding method and relevance scheme of user behavior, designed statistical parameters and DPI method proposed in this thesis lay the foundation for realizing the traffic analysis of 4G mobile network and are preliminary applied in the practical network system.