Performance Modelling of Traffic Congestion in Wireless Networks

Evolution in Wireless Technologies and Networks imposes a greater need for network support as current congestion control and avoidance techniques are mainly designed for wired networks. The current performance evaluation techniques proposed for wireless networks are not able to achieve optimal performance to guarantee desired Quality of Service (QoS) standards. Thus, the new schemes such as Active Queue Management (AQM) are needed to be adaptive to dynamic wireless networks and bursty traffic conditions to help in avoiding severe performance degradation in wireless environment. Thus, in this paper we developed and validated a novel approximate analytical performance model of a multiple threshold Random Early Detection (RED) congestion control mechanism based on the principle of Maximum Entropy (ME). It can be employed at the wireless gateways/base stations to regulate the buffer management and bandwidth allocation. Closed form expressions for the state and blocking probabilities have also been characterized. Numerical examples have been presented for aggregate and marginal QoS measures, which show the credibility of the ME solution and its validation against simulation.