QoS-aware end-to-end adaptive congestion detection and control for VoIP

Most techniques that researchers have proposed for congestion control in packet-switched networks take advantage of the congestion-control mechanism in Transmission Control Protocol (TCP). While these congestion-control methods are effective with networks that have heavy TCP traffic, they perform badly for real-time network traffic that uses the User Datagram Protocol (UDP), which does not include any congestion-control mechanism. Since real-time streams are not responsive to the congestion-control mechanism of TCP, many proposed congestion-control techniques attempt to curb real-time traffic by discarding data packets from real-time streams. As a result, these congestion-control methods degrade the Quality of Service (QoS) of real-time network applications such as Voice over Internet Protocol (VoIP) and video conferencing. We propose a new, adaptive, responsive, end-to-end technique to implement application-level congestion detection and control for real-time applications such as VoIP. Unlike existing methods, which rely on packet loss as a signal to reduce the transmission rate, our solution reacts to network congestion to anticipate and prevent packet loss, thus improving the QoS of applications that employ our algorithm.