Differentiating voice and data traffic using statistical properties
Ahmet Salih Buyukkayhan, Adnan Kavak, Ece Yaprak · 2013
In this paper, we determine the statistical properties of the Real Time Protocol (RTP) flows and enhance the Statistical Protocol Identification (SPID) application to differentiate voice and data traffic. We added 3 new attribute meters and generate a model database for the Session Initiation Protocol (SIP) and RTP protocols. The preliminary results with very low number of training capture seem promising. The results show that our new attributes improve the RTP flow identification recall by 15%.