Research And Implementation Of The AV Type Protocols Identification Methods In Mobile Network

With the advent of the mobile Internet era, mobile terminal is playing an increasingly significant role in people’s lives. Audio/Video(AV) application, which accounts for a large proportion of overall application usage, is one of the usage hotspots. By identifying this type of application, operators are able to collect user’s habit of watching audio/video online, and thus provide users with differentiated service. This identification may also serve the security audit of AV application in mobile Internet and contribute to a regulated Internet atmosphere.Nowadays, the protocol identification technique for traditional Internet has been relatively mature. However, research on identification of AV application in mobile Internet is just beginning to receive attention. This dissertation mainly study AV protocol identification technique for mobile internet. First, we compare the strengths and weaknesses and the scopes of application of traditional protocol identification techniques. We also look into the characteristics of AV application protocol in mobile Internet including huge quantity, frequent updates, etc. Then, we specifically introduce an characteristic-based protocol identification approach for AV application protocol in mobile Internet. Next, by analyzing AV application protocol data in mobile Internet, we obtain the characteristic of different application protocol, using which we implemented the AV application protocol identification. After developing AV application protocol identification program and analyzing a large quantity of experimental results, we introduce and implement automated characteristic extraction for AV application protocol in mobile Internet to improve the efficiency of identification. Last, by investigating experimental results of the automated characteristic extraction system, we develop an approach of preliminary judging AV application protocol using packet length to optimize and improve the automated characteristic extraction system. We further improve the efficiency of automated characteristic extraction and subsequently improve the efficiency of characteristic-based protocol identification approach for AV protocol in mobile Internet. After testing, this approach can be applied in actual production environments.