Statistical Analysis of Packet Interarrival Times in Wireless LAN

This paper is focused on the statistical analysis of packet arrival process in wireless LAN. The approach is fully empirical, and is based on large traces, captured from real wireless LAN. A variety of models for the packet interarrival times distribution are examined. In addition to Q-Q plot and Cramer-von Mises type statistics, are used to evaluate the models. Our findings, different from those of packet interarrival times obtained in wired backbone LAN calculations, indicate that general distribution are not applicable for wireless LAN packet arrival process. Another result of our analysis was that packet interarrival times are long-range dependent. Therefore, traffic models that assume independent exponentially distributed packet interarrival times may not produce reliable results if applied to performance analysis of wireless LAN.