Abstract: The Audio-Video(AV) type apps is a hot topic in app markets. Comparing with the traditional Internet protocol identification which is thoroughly studied, studying AV protocol identification in mobile network is preliminary. Carriers can provide differential services and security profiling by recognizing different apps from collected users’ online watching behaviors. In this paper, we study the protocol identification technology for AV apps in mobile network. We get application signatures through protocol payloads. The obtained signatures are used to classify protocols. By setting up a large number of experiments and developing identification programs, I proposed a new method that can automatically identify AV protocols in mobile network. In addition, the accuracy and efficiency are further improved.

Key words: Key Words: mobile Internet, video and audio apps, characteristic value, protocol identification