Research On QoS Technology For VoIP And Implementation

Voice over IP (VoIP) is a new and highly challenging technology which has been very hot in recent years. As requirement changes and technology progresses in telecommunications, the traditional PSTN has shown its drawbacks. At the same time, Packet Switched Data Networks has grown significantly in Quality of Service (QoS). VoIP will be the key technology of the next generation telecom, replacing traditional telecom techonology gradually. It can provide the services integrated with data and voice, and will make significant progress in the future.Though VoIP has more advantages than traditional telecom technology, it needs to be improved before being more widely accepted in practice. Quality of voice is an important factor that has a great influence on further progress. Network layer couldn't provide any guarantee of QoS for VoIP applications, due to IP network providing "best-effort" service only. Research on QoS control technology of application layer has drawn much attention recently. This thesis proposes a novel QoS control method —An Adaptive FEC Method for Real-Time Audio Transmission based on Packet Loss Prediction, and introduces some relevant technology in order to accomplish the study..Firstly, the thesis discusses the key technology on VoIP such as signaling, audio signal encoding/decoding, and real-time media transportation. Besides QoS measurement methods and QoS control strategies are addressed.Secondly, it demonstrates the correlation of one-way packet delay and loss in real-time audio transmission, shows that the packet loss is predictable, and then proposes a novel method to predict packet loss. In addition, it further evaluates performance of the packet loss prediction method, and then adopts the packet loss prediction method in packet-level FEC to proposes an adaptive FEC method.Thirdly, it constrcts a test platform based on NS-2 to test QoS control methods.Finally, it uses Vocal to implement a reliable real-time transport protocol, to demonstrate that the proposed adaptive FEC method is useful in practice.