A study on traffic and channel dynamics for wireless multimedia network performance analysis

Data arrival dynamics and channel dynamics play important role in the design of wireless multimedia networks. Network engineers must have good understanding of multimedia traffic characteristics as well as channel dynamics in order to efficiently provide the services that are envisioned for the future broadband wireless communication. We endeavor to study the effect of various data arrival statistics, channel dynamics, and their interactions on the queueing performance in wireless multimedia communication system on various platform.; In circuit switched network, we investigate the effect of voice call dynamics on the performance of data traffic in the cellular voice network platform. Our study identifies the average voice call holding time as dominant factor to affect data performance. We also propose Markov chain model which includes voice call activities as well as contentions.; In TDMA system, we characterize multipath fading/shadowing channel dynamics at the packet-level and present the work of measurement-based channel modeling for packet-level network queueing analysis. The second order channel statistics, i.e., channel power spectrum, is found to play an important role in the modeling of multipath fading channels. The data queueing performance is largely dependent on the interaction between the channel power spectrum and the data arrival power spectrum; whichever has more low frequency power will have more impact on queueing performance. Our study further examines the individual impact of mobile user driving patterns, and forward error correction, on the network performance. We propose the measurement-based Markov chain channel modeling technique, to match the important channel statistics for queueing system analysis, which shows an excellent agreement in queueing solutions between using the real original channel traces and using the sequences generated by the matched Markov chain models.; In CDMA system, the overall capacity is interference limited. The interference is affected by the aggregate traffic dynamics of users as well channel dynamics after the power control. We study the statistical multiplexing gain by the buffering in relation to traffic dynamics and power control accuracy. Further, we propose the transmission control algorithms which assign bandwidth of each user dynamically for maximum system capacity in MC-CDMA system, utilizing characteristics of MC-CDMA where interference is not linearly increasing with the rate.