Internet Telephony or Voice over Internet Protocol (VoIP) applications like Microsoft Netmeeting, Netscape Cooltalk, which use the technique of digital transmission of voice, is becoming popular. The Internet, which is currently using the best-effort (BE) service model to send packets, is not efficient enough for the distributed multimedia application like VoIP application which needs strong real-time requirements. Differentiated services (DiffServ) is a new approach proposed by IETF to provide end-to-end quality of service (QoS) for distributed multimedia application in the Internet. Expedited forwarding (EF) per-hop-behavior (PHB) in DiffServ can provide a low loss, low delay, low jitter, and assured bandwidth end-to-end service. In the thesis, the performance of VoIP over DiffServ un...[ Read more ]

Internet Telephony or Voice over Internet Protocol (VoIP) applications like Microsoft Netmeeting, Netscape Cooltalk, which use the technique of digital transmission of voice, is becoming popular. The Internet, which is currently using the best-effort (BE) service model to send packets, is not efficient enough for the distributed multimedia application like VoIP application which needs strong real-time requirements. Differentiated services (DiffServ) is a new approach proposed by IETF to provide end-to-end quality of service (QoS) for distributed multimedia application in the Internet. Expedited forwarding (EF) per-hop-behavior (PHB) in DiffServ can provide a low loss, low delay, low jitter, and assured bandwidth end-to-end service. In the thesis, the performance of VoIP over DiffServ under different scenarios is first studied. In the studies, it has been found that both delay and jitter performances of voice packets are improved when the packets are switched from BE to EF. In addition, the performances of packets in EF flows become better when excess bandwidth is given to an EF aggregate. Therefore, a scheme using Real time control protocol (RTCP) as a feedback mechanism to maintain voice quality is proposed. This helps to minimize the number of EF-marked flows in an EF aggregate to simulate the situation of excess bandwidth while the voice quality is maintained in the mean time. Moreover, the scheme is cost effective.