Discrete-time modeling for performance analysis of real-time services in IEEE 802.16 networks

摘要

The IEEE 802.16 standard has been proposed to provide last-mile wireless broadband access, especially for real-time applications. Both unsolicited grant service (UGS) and real-time polling service (rtPS) are defined to support real-time data transmission. The most important parameters of UGS and rtPS are allocated bandwidth and number of request opportunities, respectively. The critical problems of real-time applications, including latency, packet loss, and bandwidth utilization, would be impacted by different parameters configurations of UGS and rtPS. Even though the discrete-time queueing systems could be more accurate and effect in modeling frame-based systems, discrete model is not widely applied due to its complicated analysis. It motivates us to utilize a discrete-time GI-D-c model and a GI-Geo-1 model to investigate the performance of UGS and rtPS, respectively. In both models, the arrival process and service process are both derived while accounting for the traffic load, data size, and number of request opportunities. Numerous experiments indicate that the simulation results agree well with the analysis results. The UGS performs stably in most circumstances. However, latency and transmission queue size degenerate as traffic load increases when the allocated bandwidth is less than the number of arrival PDUs. The latency of rtPS is higher than the latency of UGS due to the request/grant scheme, and decreases as the number of request opportunities increases. Moreover, the bandwidth utilization of UGS and request opportunity efficiency of rtPS are also analyzed under various traffic load and linearly increase as traffic load increases. The study provides a tool to analyze the impact of different parameter configurations and the tool can be easily applicable to other scheduling services.