Skip to Main Content
The random access scheme has been shown to be an efficient transfer mechanism of packet data in wireless environments for nonreal-time applications where messages with variable length are transmitted over power controlled code-division multiple-access (CDMA) channels. Most of the previous work in this area covered the analysis of the CDMA systems for fixed or exponential packet length, infinite population, and unlimited waiting and service time where the results mainly depend on the mean values of the traffic. In this paper, we remove these assumptions and analyze the behavior of the system for the more general and realistic case of finite population, finite sojourn time, and general packet length distribution. Specifically, we provide an analytical method to study the performance related parameters of the system, such as packet delay, packet loss, and throughput as well as the effect of packet length distribution on the system performance under a realistic environment. The obtained results demonstrate that for the packet length distributions with the same mean but different tail properties, the system behavior can change dramatically. In addition, we demonstrate that this study provides an analytical tool that can be used as the underlying framework for the support of a wide range of applications and management functions such as optimization of design parameters, integration of multimedia services, and anomaly detection in CDMA wireless networks.