Voice Transmission Enhancing Model on Wireless Mesh Networks

This paper initially shows ROHC and packet aggregation significantly improve the number of successful voice calls. However, the improvement does not include processor's processing overhead, which is identified by measuring ROHC processing time from Intel Pentium 4 and RouterBOARD, and applying the results into NS-2 simulations. Simulation results indicate the number of successful voice calls is seriously affected by the processing overhead. For the solution of the processing limitation, we propose a hardware model of the two algorithms and numerically analyze the model. Simulation results show the hardware model makes the number of successful voice calls approach the ideal (no-delay) case, and the numerical model exactly characterizes the hardware behaviors.