CARP: Context-aware resource provisioning for multimedia over 4G wireless networks

Abstract

In this paper a context-aware resource provisioning framework is developed for multi-system, heterogeneous 4G wireless networks. The framework envisions that each individual sub-network is fairly independent, and uses Bayesian learning scheme for capturing user's mobility profiles. The concept of Asymptotic Equipartition Property (AEP) helps to predict the most likely path-segments, that the mobile user (MU) is going to follow in near future, with very good accuracy. This helps in proactive resource management along the MU's future paths and locations providing the optimal resource (e.g., bandwidth) reservation. Simulation results on a synthetic wireless traces corroborate this high prediction success and demonstrate sufficient improvement in delay, blocking and throughput of ongoing wireless multimedia sessions. © Springer-Verlag Berlin Heidelberg 2007.