Research on Performance Evaluation of Multimedia Traffic towards Development of Next Generation Internet

点击翻译按钮获取中文摘要

1. A queueing model with batch arrivals and random scheduling of customers where a single server takes vacations occasionally is proposed and analyzed in order to model multimedia traffic.2. QoS of communication networks is highly dependent on the frequency of packet losses, and M/G/1/K queueing model was commonly used to evaluate its performance. Concerning the scheduling of packets, FCFS was assumed so far in most cases. We have extended the model to include random scheduling and LCFS scheduling and derived the analytical results for the case.3. We have proposed CSMA/DRC which improves CSMA as a random access protocol for a variety of traffic from multimedia sources in a wireless mobile networks, and derived its performance measures.4. A mathematical method to represent self-similar and/or long-range-dependent traffic which typically appears in the Internet in terms of Wavelet transform has been developed and its usefulness has been verified.5. A method to efficiently integrate real-time traffic such as that carrying voice calls and non-real-time traffic such as that from data calls in cellular mobile networks has been developed and its performance has been evaluated.6. We have proposed a modeling technique for a class of traffic of self-similarity in terms of MMPP(Markov Modulated Poisson Process) and verified its effectiveness.7. An efficient integration scheme which accommodates both voice calls and data calls in cellular networks and which puts aside on-going data calls to an extent has been proposed and analyzed.8. The effect of long-range dependency on MPEG2 moving image traffic has been evaluated through simulation by using FGN process.