Research On Unknown Protocol Identification Model For Wireless Network

With the continuous development of modern communication technology and the explosive growth of mobile devices,wireless networks are also more and more widely used in the current communication environment.At the same time,with the popularity of personal mobile devices,wireless network security has attracted more and more attention.As the standard and specification of network communication,communication protocol has an important impact on network security.There are many communication protocols in the current wireless network environment,including a large number of private protocols.The format of private protocol is unknown,so the analysis of private protocol is very important for network security.However,due to the characteristics of communication data in wireless network,the traditional protocol identification technology can not be directly applied,so it is necessary to adopt new technology and ideas in the protocol identification,so as to solve the current problems such as low identification accuracy and large time consumption.This paper studies from two aspects of the unknown protocol identification method and identification model.The main research contents are as follows:(1)Aiming at the problem of the lack of overall design for unknown protocol identification,a model of unknown protocol identification based on self-learning is proposed.The model comprehensively considers the processes of unknown protocol identification,and reduces manual intervention by introducing the idea of self-learning.At the same time,the feedback mechanism is used in the recognition process to make it have a better recognition effect in a multi-protocol environment.Simulation results show that the model has good learning ability and high recognition accuracy.(2)In order to solve the problems of frame location accuracy and time consumption in the process of unknown protocol recognition,the idea of Text Rank,a key word extraction algorithm in natural language processing,is introduced in the process of frame location,and a method of unknown protocol frame location based on Text Rank is proposed.Using the idea of Text Rank to extract the key sequence of the unknown protocol,and using the method of multi sequence alignment to mine the synchronous sequence,as the basis of multi original data for segmentation.(3)Aiming at the problem that the accuracy and time consumption of feature extraction can't meet the actual requirements during the recognition of unknown protocols,a feature extraction method based on sequence correlation is proposed.Based on the results of frame positioning,this method divides feature extraction into two parts: intra-frame feature extraction and inter-frame feature extraction to analyze the positional relationship between feature sequences.The goal of intra-frame feature extraction is the positional relationship between different feature sequences in the same frame;the goal of inter-frame feature extraction is the positional relationship between agreed feature sequences in adjacent frames.Through the position constraints in the analysis process,the purpose of reducing time consumption and improving the accuracy of feature extraction is achieved.In summary,the model proposed in this paper can be applied to different environments,which improves the accuracy of recognition and reduces the time consumption.It has important practical and theoretical significance for the research of unknown protocol recognition in the wireless network environment.