Probability Models for Performance Evaluation and Management of Complex Communication Networks

点击翻译按钮获取中文摘要

(1) Characterization and Modeling of Multimedia TrafficWe have observed the 1/f property in the sequence of MPEG 2 variable bite rate compressed frames of a real video by the spectral analysis. We have also developed a method to generate an artificial long memory process from the sequence of the numbers of customers in an M/G/* queueing system.(2) Evaluation of Call Loss Probabilities in Cellular Mobile NetworksWe have evaluated the loss probability of new calls and the forced termination probability of hand-off calls for cellular mobile networks with non-uniform traffic conditions by numerical calculation of Markov models and simulation. Those probabilities have been compared between the current frequency division multiple access (FDMA) and the future code division multiple access (CDMA) schemes.(3) Analysis of Probability Models for Complex Communication Networks(a) We have analyzed a queueing system in which the input consists of a special semi-Markov process representing MPEG frame sequences and a Poisson process representing the interfering traffic, and evaluated the frame transmission delays based on the parameter values of real video data.(b) We have studied the throughput and delay performance of the IEEE 1394 serial interface, and compared with CSMA/CD of Ethernet by using a simulation software package QPNET.