Preference-based mobility model and the case for congestion relief in WLANs using ad hoc networks

We consider the use of campus wireless LANs (WLANs) consisting of access points (AP) with potentially noncontiguous coverage. Through surveys on mobility patterns and wireless network usage on the University of Southern California (USC) campus, we find that mobility and usage patterns exhibit significant differences across various types of locations on campus. Using the collected data, we have built a realistic mobility model, called weighted-way point (WWP), to simulate wireless network user behavior better in a university campus environment. Using WWP, we show that unbalanced wireless network usage and hotspots are likely to occur on campus. We further propose a mechanism to alleviate the local hotspot congestion by using multi-hop ad hoc networks. When a mobile node (MN) determines that the local AP is unable to provide satisfactory bandwidth to an on-going flow, it requests the flow be switched to a neighboring access point (NAP). Our mechanism differs from other schemes by allowing MNs to make the decision of initiating the flow-switching procedure and using bidirectional route discovery from both the switching MN and the NAP to reduce the route discovery delay. We use simulations to show that, with this flow-switching mechanism, flows are more evenly distributed across APs. We also show that our mechanism reduces congestion time for popular APs and improves the user-perceived quality.