The QoS-RWP Mobility and User Behavior Model for Public Area Wireless Networks

Congestion is expected to become a prominent problem to deal with as the popularity of wireless data networks continues to increase. However, this problem can in principle be mitigated if a fraction of the network users could decide to move to another location in case their perceived QoS degrades. To account for this, we propose an extension of the well-known RWP model called QoS-RWP, in which users are divided into mobile users displaying constrained movement patterns, and QoS-driven users who are mainly stationary, but they can decide to move to a better location to improve their QoS level. Another enhancement of QoS-RWP with respect to the original RWP model is that waypoints are chosen according to an access point (AP) popularity metric, which reflects the recently observed phenomenon that different APs in a wireless data network display very different degrees of popularity among users. The QoS-RWP model also accounts for different classes of load offered to the network by the users, and for different channel access methods. Based on QoS-RWP, we perform a simulationbased analysis of network usage under different combinations of network parameters such as the number of users, number of APs, relative fraction of QoS-driven users, and channel access method. Our investigation discloses interesting insights on network usage, and shows that our model is able to capture important properties observed in real-world network deployments.