Many recent studies on network traffic have demonstrated the presence of long-range dependence in local-area and wide-area network traffic. The paper empirically studies the long-range dependence properties in IEEE 802.11b wireless LAN traffic. Specifically, traffic traces in IEEE 802.11b wireless LANs are statistically analyzed, with emphasis on testing for the presence, source and degree of long-range dependence. Several different wireless network configurations are used to collect traces of network traffic. Extensive statistical analysis is then applied on the collected traffic traces to determine the presence and degree of long-range dependence. Specifically, the Hurst (H) parameter is used to measure the degree of long-range temporal dependence. The impact of different transport protocols on the degree of long-range dependence is also investigated empirically. Results of this investigation are critical for improving congestion management and quality-of-service guarantees in wireless networks.
Published in:
Computer Communications, 2003. CCW 2003. Proceedings. 2003 IEEE 18th Annual Workshop on
Date of Conference: 20-21 Oct. 2003