Research And Implementation Based On DPI Streaming Media Of Traffic Monitoring System

In recent years,the Internet’s most popular technology, P2P (Peer to Peer) streaming mediahas been fully developed. Compared with the file-sharing systems, characteristics of downloadingat the same time playing contribute to the popularity of streaming media. However, many negativeeffects arise when streaming media is universally availed, such as consuming network bandwidth,poor dissemination of information, intellectual property, information security and other issues.Therefore, further recognition and management to streaming media applications and Web trafficbecome particularly important. To be able to effectively identify and control the flow of medianetwork application traffic and prevent network users over using streaming media applications,which caused a series of problems such as lacking resources, this paper proposes a DPI-basedstreaming media traffic monitoring system. The system can effectively achieve the LANmidstream real-time identification and control of media applications. Effectively solving theseproblems can no doubt help network administrators to manage the network more securely andefficiently.Since the traffic monitoringsystem in the past can not identifythe specific of application andthe efficiency of the matching algorithm is used is too lower and the effect of limiting or blockingtraffic is not obvious defects, on the basis of several basic single-mode and multi-patternmatching algorithms studies, this paper is proposed an innovative AC-BM pattern matchingalgorithm.It has both the characteristics of multi-pattern matching and jumping-moves thought ofBM, then making it applied in traffic monitoring system identification module, combined withDPI technology depth analysis of the relevant traffic, which greatly improves system efficiency forstreaming media traffic identification; Meanwhile, the paper flow control module uses a novelcontrol method combination of straight series and bypass interference control, comparing with theprevious system of only single control method. The advantage of this approach is straightcontrolling control equipment modules deployed to some of the more easily identifiable streamingmedia applications, and some difficult structural interference packet protocol type packets areprocessed directly, which makes the pass interference part only for those are able to constructmore complex and interfering packet traffic streaming package deal, which makes the flow controlmodule controls the efficiency greatly improved. After each module system design is completed,the paper is finally on the overall functionality and performance traffic monitoring system for a comprehensive test, validated, the paper system for hundreds of species of streaming mediaapplications to accurately identify, with lower leakage identification, false identification rate,issued after the relevant control strategies. The system can effectively block the flow of streamingmedia applications, limiting the operation. The test results show that the recognition accuracy andthe control effect of the flow control system can meet the requirements of design.Currently, this paper can only identify and control on commonly used hundreds streamingmedia applications. With the development of the network and technology, streaming mediatechnology is emerging in an endless stream. However, as continuous researches are going on andsystem functions are constantly improving, the traffic monitoring system must be able to achievethe identification and control on all the streaming traffic flow. Furthermore, it can be more widelyused in network management.