Design And Implementation Of Network Traffic Control System Based On DPI

With the rapid development of the Internet, a wide range of applications in the network is emerging. People’s lives are enriched by these applications. At the same time, network bandwidth is consumed continuously. Communication content is monitored hardly. Network security problems have occurred frequently. As builders and managers of the network, the operators must monitor and control network effectively to ensure that the network run smoothly and provide different services to different customers.Traffic identification technology can classify network traffic and analyze network traffic carried by the type of application and protocol. Network bandwidth can be reasonably controlled and network threats can be prevented in time. Now a series of network traffic control products are designed and produced by use of network traffic identification technology and network traffic control technology, but most of them are designed by traditional packet inspection technology. Traditional packet inspection technology has been difficult to accurately identify the network traffic.DPI technology is a relatively simple, efficient packet inspection technology. This technology applied to traffic monitoring system will greatly improve the recognition rate and accuracy of flow on the network.Network traffic control system is designed and implemented based on operators’demand. Specifically, the work includes the following aspects:1. The mainstream identification technology and network traffic control technology is summarized by researching the status at home and abroad.2. Three DPI technology:feature identification technology, association identification technology, identification technology is introduced in detail. Scenarios of DPI technology to help operators’ network traffic identification are introduced in this paper3. The traffic control system has been overall designed. The network deployment diagram, core module design of the subsystem and performance indicators is introduced in this paper.4. A new classification of network applications is designed in this paper. The program broke the traditional ideas designed only for the application of broad categories. Innovative applications on the network are divided into six major categories. The applications are classified by application scenarios. It makes traffic identification more refined, and control more precise.5. The process and the implementation code of the traffic identification module are introduced in this paper. The process of identification based on the feature protocol, the VoIP protocol and the peer-to-peer protocol is analyzed in detail.6. Traffic management and traffic statistics module is designed and implemented. Traffic management module is designed to manage traffic and implement strategy. Traffic statistics module provides real-time traffic monitoring and statements.7. Through the test, the system achieved processing performance, network delay and detection accuracy satisfactorily.