MODELING PERFORMANCE OF VOIP TRAFFIC OVER 802.11WIRELESS MESH NETWORK UNDER CORRELATED INTER-ARRIVAL TIMES

Wireless Mesh Networks (WMN) are an attractive communication paradigm for VOIP traffic due to their low cost and relative ease of deployment. In a recent study, the end to end delay of VOIP packets in a WMN were modeled using an M/D/1 queue based model. In this model, the arrival and service patterns is assumed to follow Poisson and deterministic distributions respectively. However, the long held paradigm in the communication and performance studies that voice and data traffic follow Poisson distribution and the service time deterministic is inaccurate and inefficient. In most cases the arrivals are correlated and the sizes of requests are not similar. In this study, we propose to overcome the above challenge by modeling queuing delay and average queue length of VoIP packets in WMN to capture correlated arrivals, general size distribution and multiple channels using the MMPP/G/m queue model. The numerical results obtained from the derived models show that increasing the number of mesh routers, arrival rate and load leads to increase in queuing delay and queue length. In addition, it is observed that correlated arrivals exhibits a higher queuing delay and queue length as compared to Poisson arrivals especially at high values of mesh routers, arrival rates and load.