Utility-Based Predictive Services for Adaptive Wireless Networks with Mobile Hosts

Over the past few years, there has been a growing interest for research in wireless adaptive networking and resource management to efficiently handle wireless communications between mobile hosts and base stations. Moreover, in wireless environments, the bandwidth of an ongoing multimedia flow can dynamically be adjusted, so there must be an efficient bandwidth allocation scheme to ensure quality-of-service (QoS) guarantees and high system utilization. In this paper, the attention is focused on the management of mobile services. In particular, two classes of service, i.e., mobility independent predictive (MIP) and mobility dependent predictive (MDP), have been considered, as defined in integrated service packet networks. A utility-based rate adaptation algorithm has been considered, and an admission control has been proposed, taking into account channel conditions, through a slow-fading channel model for wireless LAN 802.11b. The valued algorithm is based on a user utility function, and the admission control can use the prereservation phase among potentially visited cells from MIP hosts while only considering the bandwidth availability on current cells for MDP services. We analyzed the MIP users' mobility along coverage areas to reduce passive resource reservations on cells that users will probably never visit through the knowledge of some mobility parameters. A prediction technique is also proposed. The performances of the wireless system have been evaluated in terms of total bandwidth utilization for MIP services, average user perceived utility, and system outage probability.