Research On VoIP Traffic Identification Technology Based On DPDK

With the rapid development of Internet,the data scale of IP Voice service(Voice over IP,VoIP)expand very quickly,which is a great challenge to network space information security.Real-time monitoring of data traffic is an important means of implementing VoIP service information security.The rapid analysis of data and the classification of accurately identification is the key of real-time monitoring of traffic.Especially for the high-speed backbone network,efficient VoIP traffic recognition is essential for network management functions.The performance of network flow recognition depends on the efficiency and accuracy of traffic identification method.At the same time,it is partially determined by receiving performance of network traffic identification within the server.The research of this article is studied from two aspects.In terms of traffic identification accuracy and efficiency,the paper introduces the statistical characteristics of VoIP traffic flow recognition algorithm based on the characteristics of the high-speed network with VoIP traffic against the shortage of traditional Vo IP traffic identification method in high speed network environment.In terms of improving the server package performance,the paper introduces the Vo IP traffic recognition framework based on DPDK technology against the low efficiency of traditional packet processing scheme based on the kernel stack problem.Then this paper designs the VoIP traffic identification prototype system based on the technology of DPDK.The main work of this paper is shown as follows:1)Aiming at the shortcomings of the Vo IP traffic identification methods respect to accuracy and efficiency in high speed network environment.Using the method of statistical analysis,this paper analyzes the VoIP packets and the other packets and studies the characteristics of VoIP traffic flow such as packets rate and packet size under high-speed networks.And then the paper put forward a kind of VoIP traffic identification method based on the character-istics of flow.The main idea is packet matching rules.The judgment for VoIP packets must comply with the rules.Finally,experiments have been carried out to verify the method.Theoretical analysis and experimental results show that VoIP traffic flow recognition algorithm based on the statistical characteristics has high recognition efficiency and accuracy.And it can meet the requirements of network operators to detect high-speed network VoIP traffic.At the same time this method also has strong scalability as well as good robustness.2)In view of the current high-speed network form,the performance of the receiving and sending inside the server is limited greatly.This paper studies the DPDK open source tech-nology,and proposes a VoIP traffic identification scheme based on DPDK open source technology.After the analysis and understanding of the DPDK platform,the Vo IP traffic identification prototype system are designed.Then the main processing flow and the data processing flow of the DPDK platform are designed.We Complete the development environment configuration,and design DPDK forwarding performance experiments with DPDK platform VoIP traffic identification experiment in this paper.By comparing the packet loss rate,the accuracy rate and the recognition efficiency under different scheme,we believe that the forwarding performance of DPDK and the VoIP traffic identification based on DPDK platform are verified.