Abstract-This paper presents a novel mechanism for dynamic rate control of prioritised Voice Over IP (VoIP) traffic in real time. The system uses our proposed variable bit rate speech codec called Speex, which can dynamically adjust the encoding bit rate (and hence the voice quality) based on the feedback information about the network congestion, flow priority, and the instantaneous speech properties. Our extensive NS2 simulation results along with results from ITU-T standard of speech quality evaluation tool (PESQ) show that the proposed system indeed provides highest quality speech while maximising the bandwidth utilisation and reducing the network congestion.
I. INTRODUCTIONVoice over IP (VoIP) has become one of the most popular IP based real-time communication applications in recent years. To support VoIP applications over the Internet, two conflicting requirements need to be met [1]. On one hand, applications are sensitive to delay, packet loss and bandwidth, so it is required to minimise the effect of network impairments on voice quality. On the other hand, since the Internet is a shared environment, resource utilisation needs to be controlled so that resource usage is optimized and congestion is avoided. This leads existing research to focus on achieving highest quality voice while both maximising the network resource utilisation and avoiding congestion.Recently, several researchers have focused on rate/quality control for multimedia flows on the Internet. Bolot et. al [2] have demonstrated a packet loss feedback based rate control scheme. However, since packet loss does not necessarily mean network congestion, algorithm based on that may not always be accurate. Mahlo et. al. [3] proposed a different approach to adapting bit-rate for VoIP flows based on an enhanced TCPfriendly rate control (TFRC) protocol that adapts the coding and packetisation to optimize the VoIP quality.In [4], researchers have also proposed a bit-rate control mechanism for VoIP application based on individual network parameters like packet loss and delay or on the predicted, perceived speech quality. In the proposed system the feedback information was sent via RTCP reports. Currently, the most common transport method for VoIP is the real-time transport protocol (RTP). Unfortunately, the cost in terms of overhead is large. Considering that most speech codecs use a frame size of 20 ms, the sum of the IP, UDP and RTP headers (20+8+12=40 bytes), sent 50 times per second represents 16 kbit/s. This means that typically at least half of the traffic in a VoIP conversation is headers. For many applications, we